Asset Liability Management PDF - Chapter 2: Modelling and Paradigms

Asset Liability Management Chapter 2: Modelling and paradigms 2. Modelling and paradigms This chapter covers the following learning objectives: Item Learnin...

Asset Liability Management Chapter 2: Modelling and paradigms 2. Modelling and paradigms This chapter covers the following learning objectives: Item Learning Objectives 1 Describe the overarching framework for asset-liability management 1.1 Discuss, in general, actuarial models as part of actuarial advice, their inputs and outputs, and how they are affected by professionalism and the external environment 1.2 Discuss the qualitative aspects of providing actuarial advice 1.3 Outline the elements of effective investment governance 1.4 Discuss the complexity of competing investment theories through describing Kuhn’s philosophy, the roles of exemplars, distinguishing between causation and correlation, and the Duhem–Quine thesis 2.1. Introduction This chapter outlines: a generic approach to modelling that is commonly used in actuarial work; theory from the philosophy of science that will help you understand the difficulties faced by an actuary when choosing investment assumptions; the distinction between investment governance and management; and the need to use qualitative methods as well as quantitative methods. 2.1.1. Thinking about the detail As discussed in Chapter 1, this subject requires you to think critically about the material presented to you and examination questions will expect you to develop clear, concise, and coherent arguments to support claims that you may make. Whilst much of the material in this subject was covered in the Foundation subjects, you should analyse the information rather than attempt to memorise the content. © June 2024 The Institute of Actuaries of Australia Page 3 of 38  Asset Liability Management Chapter 2: Modelling and paradigms A Foundation-level question may ask you to calculate the present value of a set of cashflows using a specific discount rate. In this subject, you are asked to think more deeply about what you are doing rather than just performing a calculation. Exercise 2.1 provides an example of how to think deeply about the material in this subject. Exercise 2.1 The following case study contains the thoughts of one of the most influential actuaries of the 20th century. Read the case study and produce a one-paragraph summary that shows you have thought about the issues raised. Case study: Extract from Frank Reddington’s 1958 Presidential Address 1 THE RATE OF INTEREST In this section I shall to a considerable extent speak in the first person singular. I have had inestimable benefit from my actuarial training but I am conscious that in some respects I have allowed it to have a narrowing effect. I can still recall some misgivings during my earliest lessons in that early subject in our examinations: Compound Interest. To express it now more lucidly perhaps than I felt it thirty years ago, what kind of meaning can we attach to the statement that the present value, at 4% interest, of payments of £1 per annum for twenty years is £13.59? 1 Frank Reddington (1953). Presidential address, Journal of the Institute of Actuaries, 85 (1953) 1–13. © June 2024 The Institute of Actuaries of Australia Page 4 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Case study: Extract from Frank Reddington’s 1958 Presidential Address 1 The mathematical equation is familiar to us all, but what physical reality does it represent? Two sets of payments separated by time may be equivalent, on certain assumptions; but they are never identities. For the statement to have physical reality the two parties must be able to invest all their income from the transaction, including any interest as it accrues, promptly and exactly at 4% throughout the whole of the 20 years. This hypothetical situation does not occur in the real world, so that the conception of compound interest, like any other scientific hypothesis, has an element of abstraction. Its meaning is, at least in part, conventional. The technique which results is useful and powerful: it enables us to compress a medley of varied payments spread over time into one gloriously simple figure. As a basis for contracts between two understanding parties it provides a fair and expressive language. But while it compresses it also suppresses. It suppresses both the individual personality of the series of payments and the underlying assumptions. The gain in compression is at the same time a loss in comprehension. One factor which unconsciously benumbed my sensitivity to the dangers was the use of the word 'interest ' in more than one sense. It may be a sort of shorthand to define the quantity of real physical interest in a specific contract, as in the phrase 'a loan at a rate of interest of 5%’. But if we value that contract at a rate of interest of say 4% then the 4% is a hypothetical conventional figure with no real counterpart. I suppose that I would always have claimed that I was awake to these differences, but continual traffic along this track deadened my awareness of the qualifications. Day in and day out, I became accustomed to use our basic tool—the compressor of a large array of transactions, varying in amount over time, into a single present-value—with considerable dexterity of hand and considerable absence of mind. I developed the habit of closing the other eye when I used this particular actuarial microscope. © June 2024 The Institute of Actuaries of Australia Page 5 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Case study: Extract from Frank Reddington’s 1958 Presidential Address 1 This habit has its consequences. I will use as an example an incident in my own past which I have tried to take much to heart. Some of you will remember the technical section on matching of assets and liabilities in my paper to the Institute in 1952. The importance of that problem had been in my mind for many years but I had turned back before the seemingly impossible algebraic complexity of the aggregate rate of interest buried within the extensive assets of a large life assurance fund. It was in a final mood, almost of anger, that I turned to grapple with the elusive enemy and found with a shock that I was wrestling with myself. The habits of twenty years lay between me and the obvious: that the amount of interest income was a real thing outside myself, whereas the rate of interest by which it was valued was a product of my own mind. While I clarified my own thoughts by dissociating the real amount of interest from the hypothetical rate of interest, the technique I put forward slips some way back into the rut by evaluating that real situation—the real assets and the real liabilities—with a hypothetical rate of interest. That is still the weak spot. I have elaborated upon this theme because it has relevance to later remarks and because of the sharp lesson it contains. The danger of the deep rut in my mind was not that it led to wrong answers but that it prevented me from seeing the right questions. A general observation from the case study is that we are trained to complete tasks during the early part of our actuarial education often without thinking too deeply about the validity of the method or the reasonableness of the assumptions. Many actuarial students accept the material presented to them as facts whereas many topics relating to investments are theories. The actuary-level education requires students to critically analyse methods, models, and assumptions. It is important to realise that professional actuarial work is about advising decision-makers on the implications of their choices, in addition to performing calculations. The decision-makers are likely to assume that you have performed the calculations correctly and will be more interested in understanding the story behind the numbers. It is therefore important to develop a detailed understanding of why a method is chosen for a problem and learn to question whether the underlying assumptions are valid, or at least justifiable. © June 2024 The Institute of Actuaries of Australia Page 6 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.2. General modelling framework Whilst there are many technical differences to consider when discussing the topics in this subject, there is an overarching framework that applies to them all. This generic modelling framework is discussed below and is repeated in the Fellowship subjects that cover life insurance, superannuation, and general insurance. Implicit in those Fellowship subjects, and in the description below, is an assumption that the modeler understands the domain in which the model is being constructed. The detail of the modelling process is covered in the Communication, Modelling, and Professionalism (CMP) subject, but that level of detail is not required for this subject. The general modelling framework as applied to investment advice is an example of the control cycle in action. The framework requires: data (e.g. past returns, liability profile of the investor); a method that defines how to assess the risk and return drivers; assumptions about the future; a model that can perform the calculations; output of the results; a comparison of what was expected and what occurred; communication of results; a feedback mechanism to update assumptions; and an overarching governance process. The above framework is used in subsequent chapters of this textbook. Many of the above elements are linked. For example, portfolio construction will require a model to capture the interaction between assets and liabilities and, therefore, potential mismatches. Figure 2.1 shows how a model is affected by assumptions, data, and methods. The figure also reflects how a model is influenced by professional requirements and an external environment. The general concept of modelling forms part of the Foundation, CMP, and Control Cycle subjects. We briefly recap below each of the influences on a model, as shown in Figure 2.1. © June 2024 The Institute of Actuaries of Australia Page 7 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Figure 2.1 – Modelling environment Model Assumptions Data Method Model Output 2.2.1. Models A model is a mathematical representation of a real-world phenomenon and as such must involve making simplifying assumptions about the real world. An important part of any model is the need to calibrate the model against the phenomena it is supposed to represent. The degree of simplification will depend on the purpose of the model. For example: a life insurance pricing model may use some broad asset assumptions (e.g. constant future returns), as a more refined asset model would not materially affect the derived prices; building a model to advocate investment in a new and complicated asset class, such as emerging market private credit, could involve multiple extensive models including; – a detailed investigation into each model’s sensitivities to the parameters selected for the assumptions; and – detailed communications on how to manage the risk exposures to help decision-makers decide whether to proceed with the proposed investment. © June 2024 The Institute of Actuaries of Australia Page 8 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Testing the accuracy of the calculations is necessary, involving both a technical review of the model and a peer review for materiality and sense checking. The CMP subject discusses the model review and documentation process. Modelling the future means that randomness must be considered. In one sense, all actuarial models are stochastic as the future is uncertain and events are random. However, many actuarial models appear to be deterministic as the model has been appropriately simplified and variations from a central set of assumptions are tested via rerunning the model with new parameters. It has become standard actuarial terminology to refer to these ‘simplified’ models as deterministic. For example, an actuary pricing a savings product that has no investment guarantees may assume a 6% pa, say, mean return on underlying assets. Note that there is no explicit assumption regarding the variance or other characteristics of the distribution of potential returns as only the mean return is used in this example. This does not suggest that the actuary is unaware that the actual returns will not be constant. On the grounds of materiality, the randomness of outcomes for this example can be captured through stress testing of the ‘deterministic’ model to show its sensitivities to the assumptions adopted. There are two complementary approaches used to test sensitivities for deterministic models. One approach is to apply a fixed percentage stress test to test the sensitivity of results to each assumption. The second approach is to apply a stress test to material assumptions in line with the uncertainty in those assumptions. Nowadays, it is much easier to build stochastic investment models that help the owner of a problem understand a much wider range of outputs. However, the materiality and ability to select future distributions need to be considered. For example, in Chapter 6 (Equities) we discuss fundamental analysis that involves modelling data from company accounts and other information. The purpose is to calculate the present value of a value of a company for an investor in order to compare it to the current market price. It would be unusual in those circumstances to build a stochastic model. Exercise 2.2 Outline three examples of where a deterministic model would be useful and three examples where a stochastic model would be useful. © June 2024 The Institute of Actuaries of Australia Page 9 of 38  Asset Liability Management Chapter 2: Modelling and paradigms All models require assumptions, parameters (values for certain assumptions) and data. The quality of the output from the model depends on the quality of the input (data and assumptions). Working life always imposes restrictions on resources (time, labour, IT systems), and the actuary needs to be clear to the user of the results if there are implications for the deliverables arising from any restrictions. The method, shown at the bottom of the funnel in Figure 2.1, refers to the particular asset and liability valuation procedure used. There are a variety of different methods that actuaries use. These are discussed for each of the three principal asset classes in Chapters 5, 6, and 7 and then combined with the liabilities in Chapter 11 (ALM) as part of constructing a portfolio. 2.2.2. Assumptions Assumptions relating to the drivers of risk and return are described in each of the asset class chapters (Chapters 5–7) and should be considered together with the market outlook assumptions discussed in Chapter 4 (Capital markets). Assumptions may serve to set a broad condition for a model (for example, government bonds will not default) or describe relationships within the model (for example, equity returns are highly correlated to price inflation) or set a specific figure to use in the model (for example, equity returns will be inflation plus 3% p.a.). The general process for selecting and deriving assumptions is: the assumptions of the model are identified through a mixture of factual knowledge (e.g. examining an Investment Product Disclosure Statement and historical data) and experience; the values ascribed to the assumptions (called parameters) are quantified as arising from statistical distributions and either a single value is selected for a deterministic model, or the full distribution is retained and used for a stochastic model; the quantification process involves considering: – the purpose of the advice; – the quality of data available; – the materiality of an assumption; and – how the future will differ from the past, especially considering the external environment and how that may change; and acknowledgement that assumptions and/or parameters (selected values) may be changed as experience is obtained. You should recognise the various stages of the control cycle in the list above. © June 2024 The Institute of Actuaries of Australia Page 10 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Similar thought processes apply to all aspects of investment advice. Sometimes, models are constructed that use implicit assumptions. For example, using inflation-linked bonds as a solution for hedging against domestic-related cash outflows linked by inflation implicitly assumes there is enough supply/liquidity and access to such instruments in the domestic market in practice. Care is therefore needed to accurately record what has been assumed. Assumptions will often be uncertain and so the way that actuaries manage this uncertainty will depend on the purpose of the advice. Example sources of uncertainty about whether assumptions are appropriate for a stated purpose are: when values for assumptions are deduced from an observed sample and not the population; and when assumptions are derived from data collected during a time period and then adjusted to apply to a future time period, where there may be uncertainty regarding both the size and direction of the adjustments. The uncertainty arising from assumptions depends on the degree of confidence required when selecting an assumption. Investment advice models for life insurance companies and retirement schemes are often chosen to represent mean outcomes, but some models require confidence at the 99.5% level. Exercise 2.3 Ask an AI tool like ChatGPT to explain the difference between an assumption and parameter in the context of modelling. 2.2.3. Data Data is critical to actuarial work and, in general, it is a professional requirement that actuaries comment on data issues. There may be specific regulations that specify the nature and format of disclosure of data. For example, in some jurisdictions investment advice regulations require disclosure on data issues, notably the period from which the data has been derived and how reflective that is of the current market outlook over which the investment advice is to be considered. © June 2024 The Institute of Actuaries of Australia Page 11 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Exercise 2.4 Outline issues that may arise if the source of investment return data and timing are not disclosed to retail investors. (‘Retail investors’ may have a particular meaning in a specific jurisdiction, and you need to comply with local legislation. For example, the Australian Corporations Act has a specific definition of ‘retail client’. For the purpose of answering this question, assume retail investors are individuals.) Data quality and quantity are discussed in the Control Cycle subject and practical techniques for assessing and improving data quality are introduced in the Data Analytics Principles subject. You should be aware of current market data, although you will not be tested on your knowledge of the government bond yield curve or on the current equity market P/E ratios, level of equity indexes, etc. However, it is difficult to gain an appreciation of whether a numerical result seems reasonable without actual experience. We will examine historical returns—not only to help justify future returns but also as a proxy for experience. 2.2.4. Professionalism All actuarial work should be completed in adherence to the local actuarial standards and expected professional behaviours. Some of these behaviours are explicitly stated and some may be implicit. A good question to ask when undertaking all actuarial work is how comfortable you would be defending your work in a court of law or in front of the public or on social media. Investment advice may be heavily regulated and require completion of accredited subjects. This subject is not aimed at any specific regulatory body and hence cannot be relied upon to satisfy regulatory requirements. In practice, you need to check your local Code of Professional Conduct. The Australian version is covered in the CMP subject. Broadly, the reporting of results should have regard to, among other things, the intended audience and in doing so be fit for purpose and mindful of how the results will be used by the intended audience. © June 2024 The Institute of Actuaries of Australia Page 12 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Another aspect of professionalism relates to the necessary competency to complete the requested work. An actuary should take reasonable steps, taking into account the nature of the task, to ensure they have appropriate knowledge and skills in the relevant area and at the level required in order to provide competent output. Exercise 2.5 You work as an enterprise risk actuary on life insurance products. Your company has a small business division that provides actuarial services to defined benefit (DB) schemes and has one actuary who signs off on reports. He has recently taken six months of paternity leave. You have been asked to transfer to the business division to provide sign-off on defined benefit ALM reports. Discuss your choices. 2.2.5. External environment The need to consider the external environment was covered in the Control Cycle subject. For example, actuarial work must fit in with applicable regulations and taxation rules. It is ill advised to ignore the external environment when discussing investments. Insurance companies can influence internal factors such as their expense costs and can influence some policyholder behaviour (e.g. phoning lapsed customers and asking them to reinstate their policy), but actual market investment returns are largely external to insurance company, or superannuation fund, decisions. © June 2024 The Institute of Actuaries of Australia Page 13 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.3. Investment paradigms A danger that naturally arises from spending approximately three years in university studying methods linked to actuarial work is the belief that there is an agreed body of knowledge that can be applied in practice. Whilst there are many areas where this is true, it is less true when discussing investments. We shall discuss in depth that there is significant disagreement about the validity of approaches and theories. We do not supply the ‘correct’ answer. Rather, by exposing you to arguments from all sides, we will help you formulate your views. There is a wonderful quote by the great philosopher John Stuart Mill2 that details why you need to understand both sides of an argument. He who knows only his own side of the case knows little of that. His reasons may be good, and no one may have been able to refute them. But if he is equally unable to refute the reasons on the opposite side, if he does not so much as know what they are, he has no ground for preferring either opinion... Nor is it enough that he should hear the opinions of adversaries from his own teachers, presented as they state them, and accompanied by what they offer as refutations. He must be able to hear them from persons who actually believe them... he must know them in their most plausible and persuasive form. This is not a course on the philosophy of science. We can only delve very briefly into some very deep ideas, but an examination of some philosophical issues will help frame the lack of consensus of investment theories. Parts of a paper by Craig Turnbull 3 provide a much more comprehensive coverage in the context of historical actuarial work, but that paper is significantly beyond the knowledge you need for this subject and is not assessable in this subject. This section will cover: a review of the failure of the Long-Term Capital Management hedge fund; a brief look at the difference between causation and correlation; an outline of Kuhn’s philosophy (‘normal science’ and ‘paradigm shifts’); how exemplars entrench views; the Duhem–Quine thesis, which will help you understand some of the issues involved in testing theories; and an overview of the historical development of investment theories. 2 JS Mill (1858). On liberty—see Gutenburg Press for a copy. 3 Craig Turnbull (2020). Some notes on the methodology of actuarial science, https://craigturnbullfia.com/some-notes-on- the-methodology-of-actuarial-science/ © June 2024 The Institute of Actuaries of Australia Page 14 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.3.1. Long-Term Capital Management hedge fund The mathematical methods introduced in CM2 (formerly CT8) were introduced into actuarial education towards the end of the 1990s. The methods, like many types of models, rely on assumptions that may not be borne out in practice. You should regard outputs from models that involve inferences on future states as guidance on possible outcomes that need careful interpretation, especially concentrating on the validity of the assumptions. Long-Term Capital Management (LTCM) was a hedge fund that was led by well-regarded Wall Street bond traders and Nobel Memorial Prize winners. It sought to make returns from securities that were incorrectly priced, relative to each other (see Chapter 8 for a fuller discussion). The mispricing was a small fraction of the prices of the underlying securities prices and LTCM highly leveraged its balance sheet (i.e. the capital in the balance sheet was primarily financed by loans). In 1998, Russia declared that it was devaluing its currency and defaulting on its bonds. This default was unexpected as investors expected a sovereign state would not default because it could create money—see Chapter 3 (Money in the modern economy). A flight to quality assets ensued, which pushed up the price of the most liquid securities and especially those included in widely used market benchmark indices, where LTCM were short, and depressed the prices of the less liquid assets held by LTCM. The widening of the relative prices between securities held by LTCM affected their equity and forced them to sell portfolios at unfavourable prices. These extreme events were not part of their models, although similar events certainly had occurred in the past and arguably should have influenced the construction of the models used by LTCM. The Russian crisis invalidated the following critical underlying assumptions that were used in LTCM models: sovereign states will not default on bonds or currency; perfect market liquidity—the flight to quality assets was not foreseen; and returns follow the Normal distribution—price changes were rapid and well outside plausible scenarios under the Normal distribution. The moral of the story is to always question your inputs when deciding to act on outputs from models. A compounding factor with LTCM is that the assumptions were constructed from data that covered a relatively short period of five years which may not have been representative of possible outcomes. © June 2024 The Institute of Actuaries of Australia Page 15 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Task 2.1 Find and watch a YouTube video on LTCM. This is an optional task to deepen your understanding of the potential pitfalls of assumptions when modelling. 2.3.2. Causation and correlation Background Dependency between random variables means they are somehow linked. If variables are dependent, having information about one random variable that is linked to another random variable provides information, albeit potentially unquantified, about the other random variable. Dependency does not suggest that there is a linear relationship between the two risk factors. Unfortunately, the term ‘dependence’ is often mixed with one definition of correlation. Introductory statistical courses introduce the Pearson correlation coefficient: Cov(X,Y)/√[Var(X) x Var(Y)], which is a measure of the strength of linearity between two random variables. The Pearson correlation coefficient has desirable properties in that it is easy to calculate, is familiar to many stakeholders, and it is easy to extend it to calculating sums and differences of random variables. The measurement is an indication of the linearity of the relationship, but it is not always capable of evaluating the strength of the relationship. 4 The Pearson correlation coefficient, which lies between -1 and 1, has the following limitations in evaluating the strength of relationship between two or more variables: it does not capture nonlinear effects; if the knowledge of one random variable provides exact information about another random variable, then the correlation coefficient may still take any value between -1 and 1; the correlation coefficient may be quite different under monotonic transformations of the variables— for example, the correlation coefficient of X and Y may be different from the correlation coefficient of log(X) and log(Y); 4 Look up ‘Anscombe’s quartet’ for a good explanation of this concept, the detail of which is outside the scope of this syllabus. © June 2024 The Institute of Actuaries of Australia Page 16 of 38  Asset Liability Management Chapter 2: Modelling and paradigms some insurance risks are heavy tailed, with potentially infinite variances, making the Pearson correlation coefficient indeterminable; and correlations may change over time. These limitations of the Pearson correlation coefficient may be overcome with ‘rank’ correlation methods, such as the Spearman rank correlation coefficient (‘Spearman Rho’) and the Kendal rank correlation coefficient (‘Kendall’s Tau’). These are non-parametric methods that do not depend on the marginal distributions of the underlying risk variables. The rank coefficient methods are less affected by outliers compared with the Pearson method. These tests are now included in the Foundation subjects. Causation There are two main issues with the three correlation methods discussed above: they are a numerical measure of dependency and tell part of a story but do not fully characterise the dependency structure; and independent random variables have zero correlation, but if zero correlation is measured under any of the three methods it does not necessarily imply independence of the random variables. When using these correlation methods, actuaries need to be mindful of these issues in interpreting the outcomes of dependency calculations. If two random variables, A and B, say, are (linearly) correlated, then possible relationships include: A causes B; B causes A; A and B are consequences of a common cause, but do not cause each other; A causes B and B causes A; and there is no connection between A and B; the correlation is a coincidence. The phrase correlation does not imply causality is often used and it follows from the above potential relationships. Thus, examining past data may show correlation among variables but that does not mean they are causally linked. More precisely, if A causes B, three conditions must be satisfied: 1. A must precede B; 2. B occurs if and only if A occurs; and 3. There are no other causes or effects. © June 2024 The Institute of Actuaries of Australia Page 17 of 38  Asset Liability Management Chapter 2: Modelling and paradigms If we can determine a causal relationship between two quantities, then we can be confident that the relationship continues in future periods, and we have a causal law. If we cannot determine causal relationships and can only observe historic correlation, then we have an historical regularity. We may rely on the observed historical regularity to predict out-of-sample movements (e.g. future returns), but we should recognise that this may be an unreliable prediction of the future. Many regulators ensure that investment advice provided to consumers contains a warning that past returns may not be a guide to future returns. Example Many people accept the statement that shares outperform bonds over long periods as a fact. Historical records, at first sight, appear to support that statement, but it is not universally agreed. Chapter 10 (Long-term returns) will expand on this topic. We will see in this subject that many actuaries accept that there may be significant periods where shares do not outperform bonds, but there is a long-term relationship expressed as: Return on equities = real return on bonds plus inflation plus equity risk premium. We can look at the data over a long period to see if there is a statistical basis to the above relationship. Assuming data shows this relationship is true in the historical sense, then we need to ask: why did shares outperform bonds in that period; and why will that reason be valid in the future. There is no currently accepted causal reason for the level of the equity risk premium (ERP). Thus, projecting the ERP into the future is based on a historical regularity and not a causal law. © June 2024 The Institute of Actuaries of Australia Page 18 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Exercise 2.6 Your studies of Life Contingencies would have introduced you to Gompertz’s5 law of mortality. That law assumed that the force of mortality increased with age. Makeham 6 refined the law through the addition of a constant that allowed for an age-independent rate of accidental death. Perks (1932)7 generalised those laws with an upper limit to the rate of mortality. An expression 𝐵𝑒 𝜌𝑥 for Perks’s law is 𝜇𝑥 = 𝐴 +. 1+𝐶𝑒 𝜌𝑥 In practice, the differences among the three laws are not numerically material except at young or old ages, although they are clearly different from a scientific viewpoint. Is Perks’ law a law of nature or a regularity? 2.3.3. The philosophy of Thomas Kuhn It was believed for many years, or at least taught, that science advanced by progressively accumulating more knowledge. Thus, the prevailing view was that science advances by the ‘addition of new truths to the stock of old truths, or the increasing approximation of theories to the truth, and in the odd case, the correction of past errors’. This view was developed through the Enlightenment Period, a period that saw scientific development as a rational endeavour that made gradual but persistent progress towards greater knowledge. In 1962, this picture of scientific development was challenged by Thomas Kuhn. In The structure of scientific revolutions,8 one of the most influential books of the 20th century, Kuhn proposed that science does not develop in a linear manner. This was a radical shift in thinking about scientific development. In his view, science progresses in various stages: pre-science; normal science; crisis; revolution; new normal science; and new crisis. 5 B Gompertz (1825). Philosophical Transactions, 115, 513. 6 WM Makeham (1867). Journal of the Institute of Actuaries, 13, 325. 7 W Perks (1932). Journal of the Institute of Actuaries, 63, 12 8 TS Kuhn (1962). The structure of scientific revolutions. University of Chicago Press: Chicago. © June 2024 The Institute of Actuaries of Australia Page 19 of 38  Asset Liability Management Chapter 2: Modelling and paradigms The pre-science stage is characterised by a lack of agreement over fundamentals such that it is almost impossible for different participants to agree on relevant questions, let alone agree on solutions. Normal science is the day-to-day work of scientists and involves incremental improvement of accepted theories. Kuhn’s idea is that normal science operates within a paradigm. The existence of a paradigm separates normal science from pre-science. Whilst the precise meaning of paradigm is difficult to define, according to Kuhn, it is essentially ‘the entire constellation of beliefs, values, techniques and so on shared by the members of a scientific community’. For example, a paradigm will include: clearly stated fundamental laws with associate assumptions; agreed methods of applying the fundamental laws; and definitions of measurable quantities and how to measure quantities. Normal science occurs in a world view of unquestioned assumptions that govern: the type of questions that are deemed relevant to the field of study; the process by which these questions are studied; and how results are interpreted. This was a radical shift as previously it was assumed that the role of scientists was to continually challenge the dominant paradigm rather than accept it. The dominant paradigm provides a governance structure on how to raise and answer questions. If there is not a dominant paradigm, then it can be very confusing on what are relevant questions and how to solve questions and interpret the answers. Normal science progresses by solving puzzles within the framework of the paradigm. When odd results are encountered, the researcher will question whether they did something wrong—perhaps a measuring instrument needs recalibrating. The objective is to attempt to preserve the dominant paradigm. Anomalies will appear and many will be explained away as, for example, unaccounted variables. Kuhn noted that anomalies are not counter examples of the paradigm as they are often explained away. © June 2024 The Institute of Actuaries of Australia Page 20 of 38  Asset Liability Management Chapter 2: Modelling and paradigms If anomalies accumulate that directly question the validity of the fundamentals of a paradigm and cannot be explained away, then this may push science out of ‘normal science’ into a scientific crisis. It is not possible to be precise on the onset of a crisis as it depends on how long puzzles resist being solved and the number of serious anomalies arising. The existing paradigm is not immediately rejected but becomes more diffuse in its applications. Scientists start to lose confidence in the paradigm and a revolution may occur. A new paradigm, which would lead to new normal science, must emerge before the old one is rejected. Thus, science may hold onto a paradigm even though it is insufficient to explain observations because the alternative is no paradigm and a consequent lack of rules in which to raise or answer questions. The move from paradigm to another is called a paradigm shift. The process of normal science, the accumulation of anomalies, the move into a scientific crisis, and the paradigm shift are collectively labelled by Kuhn as a scientific revolution. Turnbull’s paper, which continues this argument, is reproduced below: Revolutionary science is work that challenged the accepted beliefs of the existing paradigm and ultimately could lead to the replacement of the existing paradigm with a new one. Importantly, Kuhn argued that the new paradigm may be fundamentally inconsistent with the old one. Kuhn argued that the social and cultural norms and incentives that scientists work amongst may reduce the objectivity of scientific output. Most notably, Kuhn argued that scientists may be reluctant to give up on paradigms easily, even when they have been objectively falsified. This resistance may partly be a practical matter—there is arguably little point in rejecting a theory until a new and better one has been adequately developed. More interestingly, Kuhn argued that the leaders of the scientific community would be incentivised to resist the rejection of the paradigm in which they are experts. After all, rejection of the paradigm would imply that their work and expertise were becoming irrelevant and would be superseded by that of others. Thus, paradigm-shifts may take time (a generation or longer) and may involve significant resistance and professional controversy as they occur. © June 2024 The Institute of Actuaries of Australia Page 21 of 38  Asset Liability Management Chapter 2: Modelling and paradigms The notion that scientists may become irrationally or dogmatically attached to the particular groups of theories that they were experts in was probably not an especially new or controversial one to working scientists. For example, Max Planck, the theoretical physicist most associated with the original development of quantum theory, and hence intimately familiar with revolutionary science, wrote prior to his death in 1947: An important scientific innovation rarely makes its way by gradually winning over and converting its opponents: it rarely happens that Saul becomes Paul. What does happen is that its opponents gradually die out and that the growing generation is familiarised with the idea from the beginning.159 As is so often found in the history and philosophy of science, big ideas have a long and complicated provenance. Kuhn, however, placed Planck’s common-sense view of practising scientists within a formal philosophical framework and explored its logical consequences. Kuhn further argued that when a paradigm-shift eventually does take place (which could be decades after the first falsifications of the old theory are documented), the theories of the new paradigm could be so different to the previous one as to be ‘incommensurate’ with it. This means the terms of the theories of different paradigms may not be mutually translatable—for example, ‘mass’ in Newtonian mechanics means something different to ‘mass’ in Einstein’s theory of relativity (though it may be argued that incommensurate higher-level hypotheses need not lead to incommensurate lower-level hypotheses). Incommensurate theories may not be directly comparable. It may be the case that the new paradigm provides new or better explanations for some empirical phenomena, but at the same time is not capable of explaining some of the empirical behaviours that the previous paradigm explained well. Consequently, the old theory cannot always be merely regarded as an inferior special case or approximation of the new one. © June 2024 The Institute of Actuaries of Australia Page 22 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.3.4. Exemplars Exemplars are concrete problem-solving solutions that students encounter throughout their scientific education, whether in examinations or by examples in textbooks. The role of exemplars in entrenching paradigms is important to understand. As queried by Reddington (see case study above), it is easy to fall into the trap that a method reflects reality because we are taught to use a method without having a deep understanding of the assumptions. Reddington explained how easy it is to conflate reality with a hypothetical construct. Day-to-day usage of present-value techniques accustoms the user to not question the underlying assumptions that the hypothetical interest rate is only obtainable under almost impossible conditions. The Foundation subjects aimed to introduce students to a wide variety of techniques and the historical focus has been to illustrate methods using assumptions that allow exact solutions. For example, students are often provided with an exact probability distribution (e.g. Student’s T- distribution) and are asked to calculate the moments of the distribution. This is a useful exercise to teach the basic concepts, but students may not question whether the model is an adequate, however that is defined, representation of reality. Another example, directly relevant to this subject, is the assumption that stock returns are independent, identically distributed, log-normal random variables. Students are often asked, explicitly or implicitly, to use this assumption in order to solve various problems. The danger is that students start to assume that returns do indeed follow those assumptions because they have been repeatedly asked to make that assumption. Exercise 2.7 Can you list other examples of exemplars from your Foundation studies? © June 2024 The Institute of Actuaries of Australia Page 23 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.3.5. Duhem-Quine thesis When we examine critiques of investment theories in Chapter 9, we will encounter the difficulty of isolating, and testing, a hypothesis. This section draws on ideas from the philosophy of science and is included to demonstrate that there is a body of knowledge that considers the logic of drawing conclusions from data, which is an important part of the work of actuaries. The Duhem–Quine thesis, also known as the Duhem–Quine problem, argues that no scientific hypothesis is by itself capable of making predictions. Instead, deriving predictions from the hypothesis typically requires background assumptions that several other hypotheses are correct. The collection of background assumptions and the hypothesis are called a bundle of hypotheses. A bundle of hypotheses needs to be tested against real-world observations. Collectively, we may conclude that the bundle of hypotheses is not true (i.e. it is falsified). However, the Duhem–Quine thesis says it is impossible to isolate a single hypothesis in the bundle. One solution to the dilemma is this. When there are rational reasons to accept the background assumptions as true (e.g. explanatory scientific theories together with their respective supporting evidence), then there are rational—albeit nonconclusive—reasons for thinking that the theory under test probably is wrong in at least one respect if the empirical test fails. A concept arising from the above is underdetermination—the idea that evidence available to us at a given time may be insufficient to determine what beliefs we should hold in response to it. This concept is encountered frequently in actuarial work. We are often faced with a data set that is insufficient to provide clear conclusions, yet we need to make conclusions. Exercise 2.8 Think of examples where actuaries make conclusions using limited data. Reflection Read the article ‘Underdetermination of scientific theory’9 and summarise sections 1 and 2.1. 9 K Stanford (2017). Underdetermination of scientific theory. Stanford Encyclopedia of Philosophy. © June 2024 The Institute of Actuaries of Australia Page 24 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.3.6. Historical development of investment theory Chapter 9 provides an in-depth discussion of investment theories, and this section provides a brief overview of four investment paradigms: technical analysis (19th century); fundamental analysis (from 1934); Modern Portfolio Theory (from 1952); and behavioural finance (beginning in 1979 with Prospect Theory). Technical analysis, also called charting, was invented by Charles Dow in the late 1800s. It is a method to estimate short-term movements in asset prices based on an examination of past prices and trading volumes. Its practitioners believe that interpreting historical movements leads to an understanding of future price movements. Thus, an assumption is that there is some correlation between the combination of past price movements and trading volumes with future price movements. We will discuss this topic briefly in Chapter 6 on equities, although technical analysts apply their techniques to any asset class that is affected by the forces of supply and demand. The technique does not form a main part of this subject as we concentrate on long-term returns rates rather than short-term opportunities. Fundamental analysis is a method that attempts to derive the intrinsic value of an asset by analysing the factors that could affect its price in the future. An analyst will consider a company’s published financial statements, past trends in that data, compared with companies in similar industries, as well as other external factors, to derive an estimate of a company’s intrinsic value. This method was introduced by Benjamin Graham and David Dodd through their publication in 1934 of Security analysis, a book that proved to be highly influential. A key assumption is that the current market price is not necessarily equal to the intrinsic value of future cashflows. Through examining historical information, the analyst may derive an intrinsic value that will then lead to a determination of buying, selling, or continuing to hold a stock. Chapter 6 (Equities) outlines fundamental analysis, and the Fellowship Investment subject provides a deep dive into equity analysis. © June 2024 The Institute of Actuaries of Australia Page 25 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Modern Portfolio Theory was introduced by Harry Markowitz10 in 1952. The alternative title of mean-variance analysis summaries the method as it involves a trade-off between the return on a portfolio and the variance—the ‘risk’—of a portfolio. A series of ideas were rapidly developed such as: investors were making rational decisions; the equity risk premium as the reason why investors purchased risky assets; volatility as a proxy for risk; and correlation across assets—the ‘beta’ as part of the Capital Asset Pricing Model. The topics listed above were discussed in a technical sense in CM2 (formerly CT8). The techniques were developed in response to significant academic interest in the validity of both technical analysis and fundamental analysis. Academics observed that the average investment manager did not seem to outperform the market and perhaps that is a consequence of efficient markets. You may be aware that there are different forms of market efficiency and you may remember that the semi-strong form suggests that current prices reflect all current market information. Assuming all investors act rationally, and have the relevant current knowledge, then the current price will converge to the intrinsic value of the asset. A consequence is that fundamental analysis conveys no information because the market already ‘knows’ the information. The model of rational decision-making assumes that the decision-maker has full or perfect information about alternatives. It also assumes they have the time, cognitive ability, and resources to evaluate each choice against the others. Chapter 9 critiques this paradigm through exploring various studies for and against the paradigm. Behavioural finance grew from the development of behavioural economics, which began with the publication of Prospect theory by Kahneman and Tversky.11 Traditional investment theories and models assume investors make rational decisions and behave consistently. Broadly, it is claimed that investors aim to achieve maximum utility—more is preferred to less. Behavioural finance aims to explain actual investor behaviour. Investors are subject to biases and act irrationally as a result of these biases. For example, actual investors may not optimise their decisions as they base decisions on obtaining a satisfactory outcome rather than an optimal outcome. Chapter 9 describes various biases, although the ideas of behavioural finance have not yet been developed into a clear testable theory. 10 H Markowitz (1952). Portfolio selection, Journal of Finance, vol. 7, no. 1. 11 D Kahneman and A Tversky (1979). Prospect theory: An analysis of decision under risk. Econometrica, vol. 47, no. 2. © June 2024 The Institute of Actuaries of Australia Page 26 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.4. Governance It is important to distinguish between investment governance and investment management. These are distinctly different but related functions critical to the investment management process. Governance is focused on defining the strategy, objectives, key roles and responsibilities, documenting in an investment policy, and reviewing progress towards achieving the objectives. Management, on the other hand, is concerned with the implementation and execution of the plan. An example of one view of the various components of governance and process is outlined in Figure 2.2. This subject focuses on the governance decisions to set the investment policy and the other three components are described in detail in the Fellowship Investments subject. Figure 2.2: Investment governance and management process Investment philosophy Goals and objectives Constraints, preferences, and risk tolerances Benchmarks and measures Investment Other principles (rebalancing, review frequency, etc.) Governance Documented policy Feedback Loop Develop capital market expectations Incorporate relevant financial market, economic, social, political, and sector considerations Investment Determine risk exposure Analysis Structure portfolio (strategic asset allocation or total portfolio approach) Manager and/or individual security selection Market condition modifications Implementation Execution of portfolio decisions and Execution Rebalancing Evaluate progress towards achieving objectives Attribution analysis Benchmark comparisons Measurement Reassess strategy, tools, data, and decision-making and Attribution The investment purpose, philosophy, and goals are a board-level responsibility, made with input from the investment committee and investment staff. © June 2024 The Institute of Actuaries of Australia Page 27 of 38  Asset Liability Management Chapter 2: Modelling and paradigms A general approach to investment governance structures for institutional investors consists of three layers supporting the board: investment committee; investment staff; and third-party or external resources. The supporting governance structure may include any combination of these, depending on the size and complexity of the portfolio and Board's decision to insource or outsource certain management functions. The investment committee often consists of a subset of the board of directors, together with external investment specialists who are specifically appointed for their investment expertise. In some cases, the board of directors might decide to delegate the responsibility to an internal committee consisting of staff members. In others, the board may retain all investment governance responsibilities—that is, no investment committee. The delegation of responsibilities deemed necessary to execute the investment strategy is decided at the board level. The allocations of these accountabilities will vary between organisations depending on factors such as: the size of the organisation and funds under management; the knowledge, skills, experience, and capabilities of investment staff; the amount of time investment staff can devote, especially if they have competing priorities; and the number of investment staff and the depth of the investment team, considering business continuity, costs, efficiency, and succession planning. There is a clear distinction between the role of the board (to set the investment policy items as per first box in Figure 2.2) and the role of management (to implement and manage the portfolio, assess its performance against the policy—the detailed work in the three next boxes). The board always has the overall governance responsibility which includes oversight over the entire process, setting the policy including investment objectives, and approval of key decisions, such as an investment manager appointment. © June 2024 The Institute of Actuaries of Australia Page 28 of 38  Asset Liability Management Chapter 2: Modelling and paradigms It should be noted that there is an alternative view that setting the strategic asset allocation and the investment analysis steps needed for it should be part of investment governance. It is also important to recognise that an understanding of the characteristics of the portfolio of liabilities being funded or provided for is essential for the determination of the investment objectives for the asset portfolio and for the investment policy. This is discussed in more detail in Chapter 11 (Asset Liability Management) 2.5. Qualitative aspects The Foundation subjects focus on techniques that can solve well-defined problems. Essentially, the assumptions in the modelling of the underlying problems are over-simplified to allow students to gain stylised experience of the fundamental concepts underpinning actuarial work. In practice, an actuary as a professional advisor must consider not just quantitative techniques (‘what’ and ’when’ questions) that deliver numerical answers, but also answer qualitative questions (‘why’ and ‘how’ questions). These skills are developed in Chapter 11 (ALM). Qualitative and/or subjective input is often necessary since it is often not clear whether the past events will be replicated in the future. For example: the sample data may be too small, non-existent, or not reliable enough to build a credible statistical distribution for the assumptions; and the law of large numbers may be used to show the sample mean converges to the true underlying mean, but it assumes stationarity in the ‘true’ underlying probability distribution. The stationarity assumption is often false in practical actuarial work as changes in the external environment invalidate the notion of a unique distribution. The situation is worsened in asset pricing, as discussed in the Investment Fellowship subjects, as the current economic cycle could be completely different from investment markets experienced previously. Asset liability models place a value, or potentially a distribution of possible values, on a quantity. Models assume that probabilities can be measured. © June 2024 The Institute of Actuaries of Australia Page 29 of 38  Asset Liability Management Chapter 2: Modelling and paradigms Frank Knight in 192112 proposed a distinction between risk and uncertainty. Risks were defined as applying to situations where the outcome is unknown, but the probability of the outcome is definable. Uncertainty applies when we are incapable of having sufficient information to formulate a probability (i.e. ‘unknown unknowns’13). The above definition would suggest that modelling of assets and liabilities may fall into the uncertainty category. But values must be placed on liabilities, with appropriate caveats, and we only have past data to formulate probabilities. These probability values can be updated over time as experience emerges. Similarly, investment views about where the market could potentially be moving have to be formed, and, therefore, best estimates for risk and return expectations can be projected with specified confidence bounds. Again, these projections can be updated as market conditions crystallise. Practical actuarial work requires us to produce numbers using models, but the implications of these numbers need to be carefully explained to decision-makers. You should reflect on how you would explain the methods developed throughout this subject to a non-actuarial audience. 12 F Knight (1921). Risk, uncertainty, and profit. Boston: Houghton Mifflin. 13 Donald Rumsfeld, US Secretary of Defence, 12 February 2002. © June 2024 The Institute of Actuaries of Australia Page 30 of 38  Asset Liability Management Chapter 2: Modelling and paradigms 2.6. Key learning points The examination questions will expect you to develop clear, concise, and coherent arguments to support the claims that you may make. A model is a mathematical representation of real-world problems and involves simplifying assumptions that must be calibrated against the real world. Even if results are sometimes presented as a single deterministic outcome, most actuarial problems and models are stochastic because the future is uncertain, and events are random. Data is critical to actuarial work and the actuary must consider the accuracy of data and comment on data issues. Assumptions need to be appropriate for future experience. In setting assumptions, historical data and experience and the interpretation of emerging trends are important. All actuarial work should be completed professionally and in accordance with required standards and expected behaviours. An actuary as a professional advisor must consider not just quantitative techniques and the delivery of numerical answers, but also answer qualitative questions. The implications of the numbers you produce must be carefully explained to decision-makers. Actuarial work must also consider relevant wider issues and trends in society. Correlation does not imply causality—are past observations merely a historical record or will they occur in the future? Normal science progresses by solving puzzles within the framework of the agreed paradigm. The accumulation of anomalies may push science out of ‘normal science’ into a scientific crisis. A new paradigm, which would lead to new normal science, must emerge before the old one is rejected. The move from paradigm to another is called a ‘paradigm shift’. Exemplars are concrete problem-solving solutions that students encounter throughout their scientific education and may entrench paradigms. The Duhem–Quine thesis argues that no scientific hypothesis is by itself capable of making predictions, because background assumptions are required. There are multiple paradigms in investments, suggesting that the subject is in a ‘crisis’. Investment governance is focused on defining the strategy, objectives, key roles and responsibilities, documenting the policy and reviewing progress. Investment management is concerned with the implementation and execution of the plan decided by the governance process. Risks can be defined as applying to situations where the outcome is unknown, but the probability of the outcome is definable; and uncertainty as applying where we cannot obtain sufficient information to formulate a probability. © June 2024 The Institute of Actuaries of Australia Page 31 of 38  Asset Liability Management Chapter 3: Money in the modern economy 3. Money in the modern economy This chapter covers the following learning objectives: Item Learning Objectives 2 Explain the operation and describe the characteristics of capital markets 2.1 Discuss the role of government monetary and fiscal policy 2.2 Explain how the majority of money is created in the modern economy 2.3 Explain the implications of quantitative easing 3.1. Overview 3.1.1. What this chapter covers This chapter discusses government monetary and fiscal policy, the role of a central bank, a revision of the concept of money, and how money is predominately created in a modern economy. Apart from the section on how money is created, most of the material will be familiar to students. Whilst the focus of this subject is on descriptions of asset classes and subsequent matching to liability profiles, it is useful to outline some actions of the banking system in respect of money creation and the role of the central bank in managing inflation. The sections on the concept of money and money creation require you to read two papers published by the Bank of England. The content of the first paper, ‘Money in the modern economy: An introduction’1 should be familiar to you from your undergraduate studies. The second paper: ‘Money creation in the modern economy’ 2 describes how money is predominantly created by commercial banks issuing loans. This is different from descriptions found in some textbooks. The paper discusses the role of the central bank in the UK, but the specific detail is applicable across many modern economies. 1 M McLeay and R Thomas (2014). Bank of England Quarterly Report, vol. 54, issue 1, 4–13. 2 M McLeay and R Thomas (2014). Bank of England Quarterly Report, vol. 54, issue 1, 14–27. © June 2024 The Institute of Actuaries of Australia Page 3 of 30  Asset Liability Management Chapter 3: Money in the modern economy The final section of the second paper outlines the concept of quantitative easing. The central bank influences the amount of money in the economy. It does so in normal times by setting monetary policy—through the interest rate that it pays on reserves held by commercial banks with the central bank. Central banks have attempted to raise the quantity of broad money (i.e. cash plus other assets easily converted to cash) in circulation through quantitative easing because many economies are in a low interest rate environment. Quantitative easing affects the prices and quantities of a range of assets in the economy, including money. Definitions The following definitions will assist you in reading these materials. Overnight loans are the shortest possible term for a loan, which is a form of debt security. The cash rate is the interest rate on unsecured overnight loans between banks. The short-term money market is a market that trades debt securities with terms of no more than 365 days. Short term can mean anything from 1 to 365 days. Australian Treasury notes are a short-term discount security redeemable at face value on maturity. Terms are less than 12 months. Issued by the Australian government. Other governments have equivalent short term securities. In their simplest definition, Bank Bills are a form of an IOU (similar to a cheque, which in itself is a bill of exchange) with the exception that Bank Bills have a specific date in the future when they are payable (i.e. Maturity date). Any term can be agreed for a bank bill. In Australia, Bank Bill Swap Rates indices are determined for 1 to 6 month terms, and the Bank Bill indices are managed by the ASX. Debt Securities with a longer term (also referred to as bonds or fixed interest) are traded in the debt market (also known as bond market or fixed interest market). 3.2. Government policy This section is broadly revision of material from CB2 (Business economics). 3.2.1. Supply-side and demand-side management Government policy in respect of managing its domestic economy may be broadly split into two categories: supply-side management; and © June 2024 The Institute of Actuaries of Australia Page 4 of 30  Asset Liability Management Chapter 3: Money in the modern economy demand-side management. Supply-side management attempts to manipulate the productive capacity of an economy. The idea is that the economy on its own will not maximise economic output. Thus, government policy may be created that attempts to affect the supply of goods and services. Exercise 3.1 Write down two polices that a government might undertake to boost the supply of goods. A related concept when thinking about supply-side policies is a supply shock to the economy. This is an event that influences production capacity or costs. Exercise 3.2 Outline three supply shocks to an economy. Demand-side management attempts to stimulate the demand for goods and services. Aggregate demand in an economy has four components: the consumption of goods and services; investment in tangible assets that are used to produce goods; government spending on public goods and services; and net exports. Demand shocks are events that affect the demand for goods and services. Exercise 3.3 Outline one demand shock for each component of aggregate demand. Exercise 3.4 Discuss the relative speeds of the effects of supply-side management and demand-side management on an economy. © June 2024 The Institute of Actuaries of Australia Page 5 of 30  Asset Liability Management Chapter 3: Money in the modern economy (Hint: Start with the definitions of the two concepts and then make some broad statements about the relative speeds.) 3.2.2. Monetary and fiscal policy The two main demand-side macroeconomic tools available to governments are: monetary policy; and fiscal policy. Monetary policy is directed towards activities that aim to influence the quantity of money and credit in an economy, principally through its impact on interest rates. Changes in interest rates will change investment demand. As the quantity of money changes in an economy, the percentage of assets in investors’ portfolios will change and they may rebalance their portfolios. This rebalancing may affect the prices of other asset classes. Whilst monetary policy has a short-term effect, the long-term effect is less clear. For example, many economists believe that lowering investment rates stimulates investment and consumption demand in the short term but, in the long run, this may just translate into higher prices. That is, the stimulus may have a long-term inflationary effect. A key component in monetary policy is how to balance the trade-off between economic stimulus and inflation. Exercise 3.5 Why is it difficult not to be more definitive about the long-term effects of an expansion? Fiscal policy relates to government activities focused on taxation and spending. It is the most direct way that a government can change the growth of an economy. For example, increases in government spending increases the demand for goods and services. Another example that reduces demand would be tax rises as consumers have less money in their pockets. Fiscal policy aims to influence: the allocation of resources between different sectors and economic agents; the overall level of aggregate demand and, therefore, the level of economic activity; and the redistribution of income and wealth between different segments of the population. © June 2024 The Institute of Actuaries of Australia Page 6 of 30  Asset Liability Management Chapter 3: Money in the modern economy In a democracy, where there is a separation of powers between the executive and the legislative, there are often significant lags in gaining agreement on specific fiscal policies. This is in direct contrast to monetary policy where decisions may be outsourced to an independent central bank and thus can be implemented quickly. Both monetary policy and fiscal policy affect economic activity and can be used to regulate economic growth over time. The objectives of both monetary policy and fiscal policy are typically to create a sustainable economic environment where growth is positive and stable, and inflation is under control, being relatively low but positive. There are many challenges in the economy to achieving these objectives, including natural cycles, the so-called business cycle, and how drivers in the global economy affect the domestic economy. Exercise 3.6 The natural cycle of expansion and contraction of an economy is known as the business cycle. This cycle is not regular, and it is important to note its effect on the overall investment market. Industries that are sensitive to the business cycle are called cyclical industries. Industries that have little sensitivity are known as defensive industries. Provide some examples of cyclical and defensive industries. Reflection: Inflation—What is it? Watch this 11 minute video on Australian inflation in 2022 https://www.abc.net.au/news/2022-07-26/why-is-inflation-high-can-rba-lower-it/101267022 © June 2024 The Institute of Actuaries of Australia Page 7 of 30  Asset Liability Management Chapter 3: Money in the modern economy 3.2.3. Central banks and monetary policy tools A central bank in a country, or a monetary union of countries, is an institution that usually fulfils some, or all, of the following functions: monopoly supplier of the currency (the printed banknotes); government’s banker; bankers’ bank—the lender of last resort; regulation and/or supervision of the commercial banking system (i.e. it oversees the payments system and the financial system); and responsible for the implementation of monetary policy, primarily through the setting of interest rates. The objectives of central banks are clearly related to their roles, and so there is frequent mention of objectives related to the stability of the financial system and the payments systems. The decisions of central banks are made independently of government and this independence is important for most developed economies because it allows the execution of monetary policy without influence from short-term political pressures. Some central banks are charged with doing all they can to maintain full employment and output, whereas some also have related but less tangible roles, like maintaining confidence in the financial system or promoting an understanding of the financial sector. However, there is one overarching objective that most seem to acknowledge explicitly and that is the objective of maintaining price stability. Exercise 3.7 Which role is not usually the responsibility of a central bank? a. A banker to the government and other banks. b. Responsible for setting tax rates on interest and savings. c. Acting as the lender of last resort. © June 2024 The Institute of Actuaries of Australia Page 8 of 30  Asset Liability Management Chapter 3: Money in the modern economy Central banks have various tools available to them to influence and manage monetary policy: open-market operations; the interest rate; quantitative easing; capital requirements; and reserve requirements. An open-market operation (OMO) is an activity by a central bank to increase or decrease the liquidity of the domestic currency to, or from, a bank or group of banks. Thus, an OMO potentially influences the money supply in an economy, depending on the response of the banks. Historically, an OMO involved the purchase and sale of government bonds from, and to, commercial banks and/or designated market makers to increase or reduce the amount of money in circulation. The preferred method these days is via a repo agreement—the central bank provides money to a bank for a defined period and in return the central bank receives an eligible asset from the bank. For example, when the central bank buys government bonds from a commercial bank, the transaction should lead to an increase in the quantity of money in circulation. The increase in money arises because the central bank creates money ex nihilo—literally, out of nothing—and credits its own account with the newly created money. This newly created money is then credited to the commercial bank in return for the government bond. If banks then use these surplus funds to increase lending to corporations and households, new money is created, and this, in turn, results in broad money growth expanding. The actions of the commercial banks in making loans, and the associated constraints, are discussed in depth in Section 3.4. Similarly, the central bank can sell government bonds to commercial banks. By doing this, it reduces the quantity of money in circulation and that will reduce the ability of commercial banks to make loans (i.e. create credit) to households and corporations, thus causing broad money growth to decline. The money received by the central bank is rarely in the form of banknotes and is typically an electronic transfer. The money received may then be deleted. The interest rate that a central bank publishes is the rate at which it is willing to lend money on a very short-term basis (i.e. overnight) to the commercial banks. The interest rate is set as a target rate and not a fixed number. The central bank influences the market rate to move to the target rate through its ability to lend, or borrow, money in unlimited quantities until the market rate is close to the target rate. © June 2024 The Institute of Actuaries of Australia Page 9 of 30  Asset Liability Management Chapter 3: Money in the modern economy When a central bank announces an increase in its official interest rate, commercial banks often follow and increase their base rates at the same time. Exercise 3.8 A commercial bank’s base rate is the reference rate on which it bases lending rates to all other customers. For example, large corporate clients might pay the base rate plus 50 basis points on their borrowing from a bank, while the same bank might lend money to a small corporate client at the base rate plus 200 basis points. Why would commercial banks immediately increase their base or reference rates just because the central bank’s refinancing rate has increased? Through the setting of the official interest rate, a central bank can influence the quantity of money in the money markets. More generally, the higher the official interest rate, the higher the potential penalty (interest cost) that banks will have to pay to the central bank if they run out of liquidity and need to borrow from the central bank, which, in turn, decreases their willingness to increase lending and the more likely it is that broad money growth will shrink. Quantitative easing is a tool used in low interest rate environments where conventional 3 monetary policy tools have become ineffective. It is an extension of open-market operations as it involves the central bank purchasing financial assets from the non-banking sector. This concept is discussed in Section 3.4.2. Commercial banks are often required by local regulators to operate prudently and hold a percentage of their assets as capital. The need to hold capital prevents banks from excessive lending. 3Unconventional monetary policy the use of tools other than changing the policy interest rate. For more, see the RBA fact sheet: https://www.rba.gov.au/education/resources/explainers/unconventional-monetary-policy.html © June 2024 The Institute of Actuaries of Australia Page 10 of 30  Asset Liability Management Chapter 3: Money in the modern economy The reserve requirement is a central bank regulation that sets the minimum amount of reserves that must be held by a commercial bank. If one is required, the minimum reserve is generally determined by the central bank to be no less than a specified percentage of the amount of deposit liabilities the commercial bank owes to its customers. The commercial bank’s reserves normally consist of cash owned by the bank and stored physically in the bank vault (vault cash), plus the amount of the commercial bank’s balance in that bank’s account with the central bank. A central bank can restrict money creation by setting (and raising) a reserve requirement for banks. However, this is less common in developed economies these days. Indeed, in several countries there is no reserve requirement at all, including Australia, New Zealand, the United States of America and Sweden. Instead, the regulator might require the establishment of prudential capital and reserves to ensure ongoing viability. This, however, is aimed at maintaining solvency and not necessarily managing the economic activity more broadly. An example of a short-term policy response to external events occurred in Australia in March 2020. The Australian Prudential Regulation Authority (APRA), which regulates Australian banks, temporarily relaxed prudential capital requirements in order to aid banks to continue to lend during the depressed economic conditions flowing from the 2019–20 COVID-19 outbreak.4 If a central bank were to increase the reserve requirements, a bank that was short on reserves might have to cease its lending activities until it has built up the necessary reserves, because deposits would be unlikely to rise quickly enough for the bank to build its reserves in this way. Reserve requirements are still used in many developing countries to control lending (such as in China and India) and they remain a potential policy tool for those central banks that do not currently use it. Exercise 3.9 List three ways in which central banks can influence the supply of money. 4 Note that regulated prudential capital, or liquidity reserves, or similar, may not have to be held as accounts at the central bank. Each jurisdiction will have a particular approach. © June 2024 The Institute of Actuaries of Australia Page 11 of 30  Asset Liability Management Chapter 3: Money in the modern economy 3.3. An introduction to money This section requires you to read an article that summarises information studied in your Foundation studies. Task Read the paper: ‘Money in the modern economy: An introduction’. Exercise 3.10 What are the three primary functions of money? Exercise 3.11 What characteristics does money have to possess for it to be used as a medium of exchange? Exercise 3.12 Gold has historically been used as a form of money. What are some of the drawbacks of holding savings in physical gold today? Exercise 3.13 Describe the three different forms of money. © June 2024 The Institute of Actuaries of Australia Page 12 of 30  Asset Liability Management Chapter 3: Money in the modern economy 3.4. Money creation in a modern economy This section requires you to read a related publication from the Bank of England. We have split the reading into two parts. 3.4.1. Money creation by commercial banks Task Read the paper: ‘Money creation in the modern economy’ from the beginning (labelled page 14) until the section on quantitative easing on the page labelled 20. Exercise 3.14 Your friend argues that they have just caused the central bank to print another $25,000 in banknotes as they have received a loan approval of $250,000. She argues that the central bank requires the lender to hold a reserve equal to 10% of all loans. Explain what actually happened when she successfully applied for the loan. Exercise 3.15 A common misconception is that bank deposits are assets that can be lent out and thus create more money. Explain why this is not correct. Exercise 3.16 Loan creation and deletion are the main ways that bank deposits are created and destroyed. Give examples of other ways of creating or deleting bank deposits. © June 2024 The Institute of Actuaries of Australia Page 13 of 30  Asset Liability Management Chapter 3: Money in the modern economy Exercise 3.17 Summarise the constraints on money creation by banks. Exercise 3.18 Outline three economic conditions that could potentially increase the attractiveness of money market investments to investors who have long-time horizons and seek to maximise returns. 3.4.2. Quantitative easing Task: Money creation in the modern economy Continue with the reading, from page 20 read the section on quantitative easing ‘QE—creating broad money directly with monetary policy’. Quantitative Easing (QE) has had a more direct impact on money supply and deposits than the Open Market Operations (OMO) discussed earlier (Section 3.2.3.) as OMO do not directly increase deposits. Refer to the Figure 3 Impact of QE on balance sheets on page 24 of the reading, showing the outcomes if the central bank purchases bonds from a pension fund. In this example, the commercial bank is an intermediary and holds the cash paid by the central bank to the pension fund. This cash amount is then both an asset for the commercial bank (central bank credits the commercial bank with the purchase payment amount by increasing the commercial bank’s reserves) and a liability for the bank (as the pension fund’s cash account balance increases by the purchase payment amount). Quantitative easing is conducted by a central bank by purchasing bonds from anyone, not just from commercial banks. This leads to a direct increase in both base money and broad money as there is a payment to the bond seller (e.g. the pension fund). The pension fund assets are changed from a bond into a deposit at the pension fund’s commercial bank. This increases deposits—held as assets by the pension fund and as liabilities of the commercial bank. © June 2024 The Institute of Actuaries of Australia Page 14 of 30  Asset Liability Management Chapter 3: Money in the modern economy As the central bank is prepared to purchase the bonds, usually at low yields, QE also has the effect of reducing the cost of long term money by suppressing bond yields. This may lead to a secondary effect of increasing the willingness of consumers and businesses to borrow in order to spend, leading to a further increase in the money supply, as the commercial banks lend and create deposits (as per the Bank of England paper “Money creation in the modern economy”). The extent of this secondary effect appears to have been weaker than hoped for in the period since 2008 and partly as a result of this, inflation to date has failed to reach the 2% p.a. target set by central banks in the USA, Europe, Japan, and Australia. Where a central bank seeks to reverse the QE process by selling bonds that it has previously bought via the QE process, it is said to be carrying out Quantitative Tightening (QT). Several of the world’s largest central banks carried out QE between 2008 and 2021 and in doing so accumulated large holdings of bonds – mainly those issued by governments – and in some cases became the largest holders of government bonds in their respective economies. Since 2021 some of these banks have sought to reduce their holdings of bonds – and their presence in the bond markets – via implementation of QT. This has been a much slower process than QE and the reduction in bond holdings has been relatively modest. Open market operations is where the central bank buys bonds from a commercial bank and pays for them by increasing the value of the reserve account held at the central bank by the commercial bank. There is no direct or primary increase in deposits as a result. However, this way of buying bonds will tend to reduce bond yields and the cost of long term credit, as for QE. Remember the bond owners, primarily banks, are not under any compulsion to sell their bonds. So applying basic principles, if the central bank wishes to buy the bonds, it will have to offer an attractive (higher) price to the current owners, and higher bond prices means lower bond yields. By purchasing the bonds from banks, the central bank credits the commercial banks' accounts with cash, and then pays the central bank cash rate on the balance. The commercial bank has converted one asset (a bond) into another form of asset (cash)—and now seeks another form of investment that gets a better return than cash rate. So the commercial bank may act to increase its lending business, create more loans, and thereby increase the money supply. Therefore OMO may also lead to a secondary effect of increasing the appetite for borrowing by consumers and businesses more generally, leading to an increase in loans and deposits at commercial banks, but this may not necessarily happen as there are other factors that influence the behaviour of consumers and businesses. © June 2024 The Institute of Actuaries of Australia Page 15 of 30  Asset Liability Management Chapter 3: Money in the modern economy Exercise 3.19 Explain what would happen if the central bank rate was negative. Exercise 3.20 Explain QE and its effects on the quantity of money in an economy. Exercise 3.21 Do you agree with the comments on page 24 of the reading that QE is not free money for banks? © June 2024 The Institute of Actuaries of Australia Page 16 of 30  Asset Liability Management Chapter 3: Money in the modern economy 3.5. Key learning points The performance of an economy can be influenced using a combination of monetary policy and fiscal policy. Monetary policy is aimed at central bank activities that influence the quantity of money and credit in an economy. Fiscal policy is aimed at government activities and spending and the funding of these activities through taxation. Central banks can take on multiple roles, but, most importantly, they act as lender of last resort. Central banks are limited in their ability to manipulate the supply of money as they cannot control the amount of money that households and corporations put in banks on deposit, nor can they easily control the willingness of banks to create money by expanding credit. Instead, central banks can merely influence the supply of money through initiatives such as setting the level of interest rates and hence the cost of providing credit. Inflation targeting is the most common form of monetary policy. Exchange-rate targeting is also used, particularly in developing economies, but to a lesser extent compared to inflation targeting. Money plays three important roles, namely: – acting as a medium for exchange; – providing a mechanism for storing wealth; and – providing society with a convenient unit of account. Commercial banks create money by making loans to individuals and corporations. Quantitative easing (central bank buying assets from the non-bank sector) attempts to increase aggregate demand by significantly increasing the money supply in a low interest rate environment. © June 2024 The Institute of Actuaries of Australia Page 17 of 30 4 Quarterly Bulletin 2014 Q1 Money in the modern economy: an introduction By Michael McLeay, Amar Radia and Ryland Thomas of the Bank’s Monetary Analysis Directorate.(1) Money is essential to the workings of a modern economy, but its nature has varied substantially over time. This article provides an introduction to what money is today. Money today is a type of IOU, but one that is special because everyone in the economy trusts that it will be accepted by other people in exchange for goods and services. There are three main types of money: currency, bank deposits and central bank reserves. Each represents an IOU from one sector of the economy to another. Most money in the modern economy is in the form of bank deposits, which are created by commercial banks themselves. Money in the modern economy Subsistence economy Everyone consumes whatever they themselves produce. A farmer would consume berries and a fisherman fish. Trade Money as an IOU If someone happens to want what someone Money is a special kind of IOU that is else produces and vice versa then exchange universally trusted. may be possible. urrency printed It can take the form of currency The farmer could exchange berries for fish by the central bank, or the he deposits with the fisherman. mercial people hold in their commercial bank. In addition, for thee commercial banks Need for IOUs themselves, reserves held d with the central bank But in reality, different people represent another form want different things at different of money. times. IOUs — a promise to repay someone at a later date — can overcome this problem. The fisherman may give the farmer an IOU in exchange for berries in the summer. Then, in winter, when he