1. Theory and Models of the Financial System

Rana Foroohar

© The Financial Times, 3 May 2020

I am always surprised by how linear most economic thinking is. Economists take a stand on a particular issue - free trade is either working or it isn’t; regulation is needed or is not - and then refuse to leave their silos, even when the real world turns out again to be a messy and complex place.

The profession, sadly, is not as variegated - yet. It’s true that since the 2008 financial crisis we have moved from the neoclassical notion that the invisible hand is always right to a world in which economists also consider human bias, politics and institutional realism when crafting their models. Yet there’s still a general presumption that countries, companies, markets and individuals will eventually reset to “normal”.

Linear systems and baseline reversion to equilibrium is generally assumed. And efficiency rather than resiliency is encouraged. It was also the mainstream economics taught in universities and business schools over the past 40 years. It supported the just-in-time business culture in which redundancy in supply chains was considered a waste of time and money and the free flow of capital across borders to create more economic growth was always a good thing, despite any inequality and financial fragility they might create.

Covid-19 is ripping the scales from our eyes on such assumptions. It is also driving home an important message that policymakers need to heed. If the economics profession is going to help solve the world’s biggest problems - from pandemics and climate change to deglobalisation and inequality  -  economists must stop tweaking the edges of their models and think outside the box.

Much of the debate about how to reopen countries and cities in the wake of coronavirus has, for example, been incremental: take away X amount of social distancing, and get X amount of growth, and so on. It is as if there are some easy- to-define numerical trade-offs between the two.

“These types of phenomenon don’t work like that,” says William Hynes, the head of the OECD’s New Approaches to Economic Challenges unit, set up in the wake of the financial crisis to study how to do better policymaking within complex systems. “Pandemics aren’t linear — they are exponential. When things get knocked off track, they don’t always come back to a steady state. We’re talking about complex systems.” The same goes for the environment, populism or the financial system and, of course, the global economy.

One of the most outside-the-box economic thinkers I know, Bank of England chief economist Andy Haldane, once likened Sars to the Lehman Brothers failure.

Imagining how Covid-19 will play out - and how the corporate debt crisis or unprecedented monetary and fiscal policies may unfold - will require more creative and complex thinking than we see in most mainstream economics today. Efficiency is homogenous. Profit maximisation and shareholder “value” are clear and relatively simple to understand concepts, even if they create myriad hidden risks.

“Resiliency, on the other hand, is heterodox,” as former World Trade Organization director-general Pascal Lamy put it last week at a conference on building more shockproof global systems. Building resiliency in economic systems is harder than promoting efficiency, but ultimately it may be more rewarding.

Economists who want to help solve the world’s problems should embrace variety and complexity. The economics profession, despite some strides towards diversity, is still nearly as black and white in its thinking as are the politicians that the profession advises (I say this as someone who talks regularly to both groups and constantly has to explain how I can be both a Democrat and care about the ramifications of public debt). Economists should not only talk to more peers on the other side of the ideological spectrum, but to other experts too - biologists, environmentalists, defence experts, security analysts, engineers - and even real people.

New Approaches to Economic Challenges (NAEC) https://www.oecd.org/naec/

OECD/NAEC-OMI conference: Shock-proof: Building Resilient Systems in the 21st Century https://www.oecd.org/naec/events/systemic-linkages/building-resilient-systems-in-21st-century.htm

Andrew G. Haldane

It would be easy to become very depressed at the state of economics in the current environment. Many experts, including economics experts, are simply being ignored. But the economic challenges facing us could not be greater: slowing growth, slowing productivity, increased protectionism, the retreat of globalisation, high and rising levels of inequality. These are deep and diverse problems facing our societies and we will need deep and diverse frameworks to help understand them and to set policy in response to them. In the pre-crisis environment when things were relatively stable and stationary, our existing frameworks in macroeconomics did a pretty good job of making sense of things.

But the world these days is characterised by features such as discontinuities, tipping points, multiple equilibria, and radical uncertainty. So if we are to make economics interesting and the response to the challenges adequate, we need new frameworks that can capture the complexities of modern societies.

We are seeing increased interest in using complexity theory to make sense of the dynamics of economic and financial systems. For example, epidemiological models have been used to understand and calibrate regulatory capital standards for the largest, most interconnected banks, the so-called “super-spreaders”. Less attention has been placed on using complexity theory to understand the overall architecture of public policy – how the various pieces of the policy jigsaw fit together as a whole in relation to modern economic and financial systems. These systems can be characterised as a complex, adaptive “system of systems”, a nested set of sub-systems, each one itself a complex web. The architecture of a complex system of systems means that policies with varying degrees of magnification are necessary to understand and to moderate fluctuations. It also means that taking account of interactions between these layers is important when gauging risk.

Although there is no generally-accepted definition of complexity, that proposed by Herbert Simon in The Architecture of Complexity - “one made up of a large number of parts that interact in a non-simple way” - captures well its everyday essence. The whole behaves very differently than the sum of its parts. The properties of complex systems typically give rise to irregular, and often highly non-normal, statistical distributions for these systems over time. This manifests itself as much fatter tails than a normal distribution would suggest. In other words, system-wide interactions and feedbacks generate a much higher probability of catastrophic events than Gaussian distributions would imply.

For evolutionary reasons of survival of the fittest, Simon posited that “decomposable” networks were more resilient and hence more likely to proliferate. By decomposable networks, he meant organisational structures which could be partitioned such that the resilience of the system as a whole did not rely on any one sub-element. This may be a reasonable long-run description of some real-world complex systems, but less suitable as a description of the evolution of socio-economic systems. The efficiency of many of today’s networks relies on their hyper-connectivity. There are, in the language of economics, significantly increasing returns to scale and scope in a network industry. Think of the benefits of global supply chains and global interbank networks for trade and financial risk-sharing. This provides a powerful secular incentive for non-decomposable socio-economic systems.

Moreover, if these hyper-connected networks do face systemic threat, they are often able to adapt in ways which avoid extinction. For example, the risk of social, economic or financial disorder will typically lead to an adaptation of policies to prevent systemic collapse. These adaptive policy responses may preserve otherwise-fragile socio-economic topologies. They may even further encourage the growth of connectivity and complexity of these networks. Policies to support “super-spreader” banks in a crisis for instance may encourage them to become larger and more complex. The combination of network economies and policy responses to failure means socio-economic systems may be less Darwinian, and hence decomposable, than natural and biological systems.

What public policy implications follow from this complex system of systems perspective? First, it underscores the importance of accurate data and timely mapping of each layer in the system. This is especially important when these layers are themselves complex. Granular data is needed to capture the interactions within and between these complex sub-systems.

Second, modelling of each of these layers, and their interaction with other layers, is likely to be important, both for understanding system risks and dynamics and for calibrating potential policy responses to them.

Third, in controlling these risks, something akin to the Tinbergen Rule is likely to apply: there is likely to be a need for at least as many policy instruments as there are complex sub-components of a system of systems if risk is to be monitored and managed effectively. Put differently, an under-identified complex system of systems is likely to result in a loss of control, both system-wide and for each of the layers.

In the meantime, there is a crisis in economics. For some, it is a threat. For others it is an opportunity to make a great leap forward, as Keynes did in the 1930s. But seizing this opportunity requires first a re-examination of the contours of economics and an exploration of some new pathways. Second, it is important to look at economic systems through a cross-disciplinary lens. Drawing on insights from a range of disciplines, natural as well as social sciences, can provide a different perspective on individual behaviour and system-wide dynamics.

The NAEC initiative does so, and the OECD’s willingness to consider a complexity approach puts the Organisation at the forefront of bringing economic analysis policy-making into the 21st century.

From economic crisis to crisis in economics, OECD Insights http://oecdinsights.org/2017/01/11/from-economic-crisis-to-crisis-in-economics/

NAEC Roundtable, OECD, 14 December 2016, http://video.oecd.org/players/dlmpBtVZ-zLampSRtlFiNFxaLN1Z7FtwkYN18--Gm2b2LSng85YLWggTcxKx2OqA9sgDoGK2bb5FCgme1SLoCg

“The dappled world” GLS Shackle Biennial Memorial Lecture, 10 November 2016 https://www.bankofengland.co.uk/-/media/boe/files/speech/2016/the-dappled-world.pdf

“On microscopes and telescopes”, Lorentz Centre, Leiden, Workshop on socio-economic complexity, 27 March 2015 https://www.bankofengland.co.uk/speech/2015/on-microscopes-and-telescopes

Michael Bordo

Economic development and growth in the past two centuries have been facilitated by stabilising monetary and financial regimes, in large part thanks to central banks developing policy tools to provide both macroeconomic and financial stability. Macroeconomic stability comprises price level stability (today low inflation); limited volatility in the real economy (smoothing the business cycle) and financial stability. Traditionally, financial stability meant preventing and managing financial crises but recently it has come to mean heading off systemic risk (imbalances) and especially credit-driven asset price booms and busts which can trigger financial crises. Since the Global Financial Crisis (GFC) of 2008, central banks have focused increasingly on their financial stability mandate and especially the link between credit-driven asset price booms and busts which many view as the key cause of financial crises.

The consensus among economists and policy makers is that credit driven asset price booms are the key cause of serious financial crises. The existence of such a link can be tested first by surveying the co-evolution of monetary policy and financial stability and the historical evidence on the incidence, costs and determinants of financial crises; then looking at empirical historical evidence on the relationships between credit booms, asset price booms and serious financial crises. This will help answer the question of whether the two serious financial crises which were linked to credit-driven asset price booms and busts, the 1929-33 “Great Contraction” and the GFC, should be grounds for permanent changes in the monetary and financial environment.

To provide some empirical perspective, I examined the evidence for a sample of 15 advanced countries from 1880 to the present to see if credit booms associated with banking crises peak slightly before or are coincident with banking crises; if equity boom busts and housing price boom busts associated with banking crises occur shortly before or coincident with serious banking crises; and to determine the relationship between these types of events and banking crises associated with severe recessions.

The results suggest that credit boom induced big crises like the Great Contraction or the GFC are very rare - about once in every 50 years - and that credit booms are not very closely connected to asset price booms. Credit driven asset price booms were important in a few big crises before World War II but not the majority. Financial instability has though returned with the return of financial globalisation since the collapse of Bretton Woods and the liberalisation of domestic financial sectors. Since the 1970s major financial innovation has allowed banks to fund themselves in the financial markets and not have to rely on deposits. This has allowed bank credit to grow faster than the money supply, has increased leverage, and may have been a key factor triggering asset price booms and possible financial crises since the 1980s.

In addition, financial innovation, made possible by the growth of financial theory and financial innovation, has led to the growth of non-bank financial intermediaries (shadow banks) which are outside the traditional supervisory and regulatory networks. These innovations both in the traditional banking sector and the shadow banking sector have increased both leverage and liquidity in the financial system. This has created a new source of systemic risk which can increase financial instability.

Moreover, output losses in the period since 1997 are much larger than in the pre-1914 period despite today’s greater reliance on lender of last resort policies and other policies designed to remedy the market failures associated with financial shocks. This may be explained by the fact that in recent years deposit insurance and the financial sector safety net created guarantees of the financial system which converted banking panics into fiscally resolved financial crises which became increasingly more expensive to resolve. The stakes associated with financial crises have therefore been higher, reinforcing the imperative for monetary authorities to prevent them.

The means to do so should however be considered carefully before advocating radical reform. After the Great Contraction the world’s monetary authorities believed that a sea change in monetary policy and financial stability policy was called for, and repressed both the domestic and international financial system for 40 years. That strategy led to unintended consequences driven by the dynamics of financial innovation and may in turn have sown the seeds for the GFC 80 years later. An obsession with financial stability (and the increased use of the tools of macroprudential policy and “leaning against the wind” by using monetary policy tools to head off imbalances) raises the risk of repeating the mistakes of the 1930s and creating a new regime of financial repression which will also have unintended consequences.

The analogy with policies designed to suppress natural disasters should be kept in mind, especially the trade-off between efficiency and resilience. Extinguishing every small fire in a forest may seem like a useful precaution, but it allows undergrowth to proliferate, adding potential fuel to future fires, just as dampening volatility in financial markets encourages risk-taking and increases the chances of a crisis.

A key lesson from the historical record through the Great Moderation period is that if four key principles are followed a stable monetary policy regime can be compatible with financial stability: price stability (credibility for low inflation); real macro stability (via e.g. flexible inflation targeting); a credible rules-based lender of last resort, and sound financial supervision and regulation and banking structure. Canada, which followed these principles, avoided banking crises altogether.

The GFC and the Great Recession were contained by effective monetary and fiscal policies and an unorthodox extension of the lender of last resort by the Fed and other authorities who had learned the lessons of the 1930s. However, like the 1930s, the GFC was blamed on the banks and the financial system and this has led to the creation of a new regime of financial regulation and the elevation of the financial stability mandate to primary importance. In addition to financial repression, the adoption of many of the tools of macroprudential regulation that have been proposed may recreate the problems with the use of these tools in the past.

Many of these macroprudential policies were actually credit or fiscal policies which greatly involved the monetary authorities in inefficiently picking winners and losers and influencing the allocation of resources. They also harmed central bank independence because these policies strayed from their mandates and opened them up to scrutiny and criticism by the legislature. An enhanced financial stability strategy may put out some small fires in the coming years, but precipitate an even bigger crisis than 2008 a few decades from now.

A knowledge of history matters. Basing important regime-changing decisions on the last crisis ignores the heterogeneity of crises. History teaches us the importance of relearning the details of the events of the past which often contain important and long forgotten clues to aid in our understanding of a current crisis.

NAEC seminar “The Great Financial Crisis and the Recovery”, OECD, 4 September 2018 http://www.oecd.org/naec/events/great-financial-crisis-and-recovery.htm

“An historical perspective on the quest for financial stability and the monetary policy regime” NBER Working Paper 24154, December 2017 http://www.nber.org/papers/w24154

Andrew Lo

Economic behaviour and financial markets are a product of human evolution, and as such are shaped by biological laws. The basic principles of mutation, competition, and natural selection apply to the banking industry as much as to natural ecosystems. The key to these laws is adaptive behaviour in shifting environments. To understand the complexity of human behaviour, we need to understand the different environments that have shaped it over time and across circumstances. We need to understand how the financial system functions and sometimes fails under these different conditions. We have assumed rational economic behaviour for so long that we’ve forgotten about other aspects of human behaviour.

Neuroscience and evolutionary biology confirm that rational expectations and the Efficient Markets Hypothesis (EMH) capture only some of the full range of human behaviour. That is not to say we should discard EMH altogether. It takes a theory to beat a theory, and the behavioural finance literature has yet to offer a clear alternative that does better. Psychology, neuroscience, evolutionary biology, and artificial intelligence can all help us to understand market behaviour, but none of them offers a complete solution. We need a new narrative for how markets work, and now have enough pieces of the puzzle to start putting it all together.

We begin by acknowledging that market inefficiencies exist. These inefficiencies and the behavioural biases that create them are important clues into how the brain makes financial decisions. We’ve seen how biofeedback measurements can be used to study behaviour, and can use imaging techniques to watch how the human brain functions in real time as we make decisions. However, neuroeconomics is only one layer. For example, neuroscience can tell us why people with dopamine dysregulation syndrome become addicted to gambling, but it doesn’t explain anything about the larger picture of financial decision making. To the sceptic, the peculiar behaviours described in these neuroscientific case studies are really just “bugs” in the basic program of economic rationality, the exceptions that prove the rule.

In fact, we have to turn the standard economic view of human rationality on its head. We aren’t rational actors with a few quirks in our behaviour. Our brains are collections of quirks. Working together, under certain conditions, these quirks often produce behaviour that an economist would call “rational.” But under other conditions, they produce behaviours that an economist would consider wildly irrational. These quirks are the products of brain structures whose main purpose isn’t economic rationality, but survival.

Our neuroanatomy has been shaped by the long process of evolution, changing only slowly over millions of generations. Our behaviours are shaped by our brains. Some of our behaviours are evolutionarily old and very powerful. The raw forces of natural selection, reproductive success or failure— in other words, life or death— have engraved those behaviours into our very DNA. Natural selection gave us abstract thought, language, and the memory-prediction framework. These adaptations give us the power to change our behaviour within a single lifespan, in response to immediate environmental challenges and the anticipation of new challenges. Natural selection also gave us heuristics, cognitive shortcuts, behavioural biases, and other conscious and unconscious rules of thumb— the adaptations that we make at the speed of thought. Natural selection isn’t interested in exact solutions and optimal behaviour, features of Homo economicus. Natural selection only cares about differential reproduction and elimination, in other words, life or death. Our behavioural adaptations reflect this cold logic. However, evolution at the speed of thought is far more efficient and powerful than evolution at the speed of biological reproduction, which unfolds one generation at a time. Evolution at the speed of thought allows us to adapt our brain functions across time and under myriad circumstances to generate behaviours that have greatly improved our chances for survival.

This is the core of the Adaptive Markets Hypothesis, whose basic idea can be summarised in five key principles:

  1. 1. We are neither always rational nor irrational, but we are biological entities whose features and behaviours are shaped by the forces of evolution.

  2. 2. We display behavioural biases and make apparently suboptimal decisions, but we can learn from past experience and revise our heuristics in response to negative feedback.

  3. 3. We have the capacity for abstract thinking, specifically forward-looking what- if analysis; predictions about the future based on experience; and preparation for changes in our environment. This is evolution at the speed of thought, which is different from but related to biological evolution.

  4. 4. Financial market dynamics are driven by our interactions as we behave, learn, and adapt to each other, and to the social, cultural, political, economic, and natural environments in which we live.

  5. 5. Survival is the ultimate force driving competition, innovation, and adaptation.

Under the Adaptive Markets Hypothesis, individuals never know for sure whether their current heuristic is “good enough.” They make choices based on their experience and their best guess as to what might be optimal. They learn by receiving positive or negative reinforcement from the outcomes. As a result of this feedback, individuals will develop new heuristics and mental rules of thumb to help them solve their various economic challenges. As long as those challenges remain stable over time, their heuristics will eventually adapt to yield approximately optimal solutions to those challenges.

The Adaptive Markets Hypothesis can easily explain economic behaviour that is only approximately rational, or that misses rationality narrowly. But it can also explain economic behaviour that looks completely irrational, as when the environment changes and the heuristics of the old environment might not be suited to the new one. Or when individuals receive no reinforcement from their environment, and don’t learn. Likewise, inappropriate reinforcement will teach individuals suboptimal behaviour. And if the environment is constantly shifting, individuals may never reach an optimal heuristic. This, too, will look “irrational.”

The Adaptive Markets Hypothesis recognises that suboptimal behaviour is going to happen when we take heuristics out of the environmental context for which they emerged. Even when an economic behaviour appears extremely irrational, it may still have an adaptive explanation. Such behaviour isn’t “irrational,” but “maladaptive.” Our behaviour adapts to new environments both in the short term as well as across evolutionary time, and not always in financially beneficial ways. Financial behaviour that may seem irrational now is behaviour that has not had sufficient time to adapt to the context. Economic expansions and contractions are the consequences of individuals and institutions adapting to changing financial environments, and bubbles and crashes are the result when the change occurs too quickly.

NAEC seminar “New approaches to financial markets”, OECD, 20 October 2017, https://youtu.be/MW6zqaKP-dU

Adaptive Markets: Financial Evolution at the Speed of Thought, Andrew W. Lo, Princeton University Press, 2019

Richard Bookstaber

Traditional economics cannot address well four characteristics of human experience that manifest themselves in crises. The first of these “Four Horsemen of the Econopalypse” is computational irreducibility. You may be able to reduce the behaviour of a simple system to a mathematical description that provides a shortcut to predicting its future behaviour, the way a map shows that following a road gets you to a town without having to physically travel the road first. Unfortunately, for many systems you only know what is going to happen by faithfully reproducing the path the system takes to its end point, through simulation and observation, with no chance of getting to the final state before the system itself. It’s a bit like the map Borges describes in On Rigor in Science, where “the Map of the Empire had the size of the Empire itself and coincided with it point by point”. Not being able to reduce the economy to a computation means you can’t predict it using analytical methods, but economics requires that you can.

The second characteristic property is emergence. Emergent phenomena occur when the overall effect of individuals’ actions is qualitatively different from what each of the individuals are doing. You cannot anticipate the outcome for the whole system on the basis of the actions of its individual members because the large system will show properties its individual members do not have. For example, some people pushing others in a crowd may lead to nothing or it may lead to a stampede with people getting crushed, despite nobody wanting this or acting intentionally to produce it. Likewise no one decides to precipitate a financial crisis, and indeed at the level of the individual firms, decisions generally are made to take prudent action to avoid the costly effects of a crisis. But what is locally stable can become globally unstable.

The third characteristic “non-ergodicity”, comes from German physicist Ludwig Boltzmann who defined as “ergodic” a concept in statistical mechanics whereby a single trajectory, continued long enough at constant energy, would be representative of an isolated system as a whole, from the Greek “ergon”, energy, and “odos”, path. The mechanical processes that drive of our physical world are ergodic, as are many biological processes. We can predict how a ball will move when struck without knowing how it got into its present position – past doesn’t matter. But the past matters in social processes and you cannot simply extrapolate it to know the future. The dynamics of a financial crisis are not reflected in the pre-crisis period for instance because financial markets are constantly innovating, so the future may look nothing like the past.

Radical uncertainty completes our quartet. It describes surprises—outcomes or events that are unanticipated, that cannot be put into a probability distribution because they are outside our list of things that might occur. Electric power or the internet are examples from the past, and of course we don’t know what the future will be. As Keynes put it, “There is no scientific basis to form any calculable probability whatever. We simply do not know.” Economists also talk about “Knightian uncertainty”, after Frank Knight, who distinguished between risk, for example gambling in a casino where we don’t know the outcome but can calculate the odds; and “true uncertainty” where we can’t know everything that would be needed to calculate the odds. This in fact is the human condition. We don’t know where we are going, and we don’t know who we will be when we get there. The reality of humanity means that a mechanistic approach to economics will fail.

So is there any hope of understanding what’s happening in our irreducible, emergent, non-ergodic, radically uncertain economy? Yes, if we use methods that are more robust, that are not embedded in the standard rational expectations, optimisation mode of economics. To deal with crises, we need methods that deal with computational irreducibility; recognise emergence; allow for the fact that not even the present is reflected in the past, never mind the future; and that can deal with radical uncertainty. Agent-based modelling could be a step in the right direction.

Agent-based models (ABM) use a dynamic system of interacting, autonomous agents to allow macroscopic behaviour to emerge from microscopic rules. The models specify rules that dictate how agents act based on various inputs. Each agent individually assesses its situation and makes decisions on the basis of its rules. Starlings swirling in the sky is a good illustration. The birds appear to operate as a system, yet the flight is based on the decisions of the individual birds. Building a macro, top-down model will miss the reality of the situation, because at the macro level the movements of the flock are complex, non-linear, yet are not based on any system-wide programme. But you can model the murmuration based on simple rules as to how a bird reacts to the distance, speed and direction of the other birds, and heads for the perceived centre of the flock in its immediate neighbourhood.

Likewise, the agent-based approach recognises that individuals interact and thereby change the environment, leading to the next interaction. It operates without the fiction of a representative consumer or investor who is as unerringly right as a mathematical model can dream. It allows for construction of a narrative—unique to the particular circumstances in the real world—in which the system may jump the tracks and careen down the mountainside. This narrative gives us a shot at pulling the system back safely.

In short, agent-based economics arrives ready to face the real world, the world that is amplified and distorted during times of crisis. This is a new paradigm rooted in pragmatism and in the complexities of being human.

For the financial system, we model liquidity, leverage and concentration. We plug in values we believe to accurately represent the current state of economic market. We run simulations to see how often things go off the rails. If things don’t go off the rails often we make it green, if not we make it red. If you increase both liquidity and leverage, things tend to get worse. When we model risk in time according to static models, scenarios of crisis and boom are equally likely. When we apply ABM, crisis and boom are not symmetrical. Crisis is not a single bad draw from a homogeneous distribution of risks. It causes a cascade.

In the analytic deductive approach, you plan everything from start to finish and then fill it all in. This is not the way to approach a crisis where the unexpected will always happen. A better way is the “headlights on the road” approach. You go ahead and see where the next curve is. Once you get there, you see where the following curve is, and so on. You solve it as far as you can see, and you are always adapting. In a crisis, you should be able to make changes, test critical assumptions and variables.

NAEC seminar “The end of theory”, OECD, 29 June 2017, http://www.oecd.org/naec/events/the-end-of-theory-financial-crises-failure-of-economics-and-sweep-of-human-interaction.htm

“Agent-based models to help economics do a better job”, Richard Bookstaber, OECD Insights, http://oecdinsights.org/2017/01/23/agent-based-models-to-help-economics-do-a-better-job/

Jean-Philippe Bouchaud

The crisis put classical economics under pressure. In theory, deregulated markets should be efficient, with rational agents quickly correcting any mispricing or forecasting error. Prices should reflect the underlying reality and ensure optimal allocation of resources. These “equilibrated” markets should be stable: crises can only be triggered by acute exogenous disturbances not the market itself. This is in stark contrast with most financial crashes.

The crisis might offer an occasion for a paradigm change, to which physics could contribute, through so-called econophysics. Econophysics has tended to concentrate on financial markets, and these represent an ideal laboratory for testing economics concepts using the terabytes of data generated every day by financial markets to compare theories with observations.

In financial markets, physicists are intrigued by a number of phenomena described by power-laws. For example, the distribution of price changes, of company sizes, of individual wealth all have a power-law tail, to a large extent universal. The activity and volatility of markets have a power-law correlation in time, reflecting their intermittent nature, obvious to the naked eye. Many complex physical systems display very similar intermittent dynamics, for example velocity fluctuations in turbulent flows. While the exogenous driving force is regular and steady, the resulting endogenous dynamics is complex and jittery. In these cases, the non-trivial (physicists say “critical”) nature of the dynamics comes from collective effects: individual components have a relatively simple behaviour, but interactions lead to new, emergent phenomena. The whole is fundamentally different from any of its sub-parts. The dynamics of financial markets, and more generally of economic systems, may reflect the same underlying mechanisms.

Several economically-inspired models exhibit these critical features. One (a transposition of the Random Field Ising Model, RFIM) describes situations where there is a conflict between personal opinions, public information, and social pressure. Traders are influenced by some slowly varying global factors, for example interest rates or dividend forecasts. Assume no shocks in the dynamics of these exogenous factors, but that each trader is influenced by the opinion of the majority. If all agents made up their mind in isolation (zero herding tendency) then the aggregate opinion would faithfully track the external influences and, by assumption, evolve smoothly.

But if the herding tendency exceeds some finite threshold, the evolution of the aggregate opinion jumps discontinuously from optimistic to pessimistic, while global factors only deteriorate slowly and smoothly. Furthermore, some hysteresis appears. Like supersaturated vapour refusing to turn into liquid, optimism is self-consistently maintained. To trigger the crash, global factors have to degrade far beyond the point where pessimism should prevail. Likewise, these factors must improve much beyond the crash tipping point for global optimism to be reinstalled.

The representative agent theory amounts to replacing an ensemble of heterogeneous and interacting agents by a unique representative one, but in the RFIM, this is impossible: the behaviour of the crowd is fundamentally different from that of any single individual.

Minority Games define another, much richer, family of models in which agents learn to compete for scarce resources. A crucial aspect here is that the decisions of these agents impact the market: the price does not evolve exogenously but moves as a result of these decisions. A remarkable result here is the existence of a phase transition as the number of speculators increases, between a predictable market where agents can make some profit from their strategies, and an over-crowded market, where these profits vanish or become too risky.

There are other examples in physics and computer science where competition and heterogeneities lead to interesting phenomena, for example cases where even if an equilibrium state exists in theory, it may be totally irrelevant in practice, because the equilibration time is far too long.

As models become more realistic, analytics often has to give way to numerical simulations. This is well-accepted in physics, but many economists are still reluctant to recognise that numerical investigation of a model, although very far from theorem proving, is a valid way to do science. It is surprising how easily numerical experiments allow one to qualify an agent-based model (ABM) as potentially realistic or completely off the mark. What makes this expeditious diagnosis possible is the fact that for large systems details do not matter much – only a few microscopic features end up surviving at the macro scale.

The attraction of ABM is that they can put together simple elements that produce rich behaviours. The instability mechanisms in the complex systems they are used to study show common features. Phase diagrams are a core element of this approach, allowing the study of places where behaviour can change suddenly and radically. In ABM, macro observables such as output are not smooth functions of the parameters. The interest rate for example can induce a transition between a good and a bad phase.

The notion of emergence is important in ABM. Equilibrium output level is usually exogenous in traditional, models, but in an ABM it is the result of the ability of agents (or firms) to co-operate, so it is an emergent property that can appear or disappear suddenly. This is one way to think about crises.

ABM also allow for the notion of hysteresis. Different states of the economy can coexist in the same region of parameter space. The economy can be stuck in a good or a bad state, while the system could have chosen another outcome if the history had been different or some anecdotal event occurred.

ABM can allow policy experiments, even if they still require a lot of work as policy tools. They show that policies that would be stabilising if you assume infinitely rational, forward-looking agents can actually be destabilising when you remove that assumption. This makes it intrinsically difficult to design hybrid models incorporating some elements of ABM. If you abandon infinitely forward-looking agents, things happen that would not happen with them. In addition, there is the “curse of complexity”. Optimised complex systems are often on the verge of instability – optimality and instability go hand in hand.

Other empirical results, useful analytical methods and numerical tricks have been established by econophysics, which I have no space to review here, but the most valuable contribution may be methodological nature. Physics constructs models of reality based on a subtle mixture of intuition, analogies and mathematical spin, where the ill-defined concept of plausibility can be more relevant than the accuracy of the prediction. Kepler’s ellipses and Newton’s gravitation were more plausible than Ptolemy’s epicycles, even when the latter theory, after centuries of fixes and stitches, was initially more accurate to describe observations. Physicists definitely want to know what an equation means in intuitive terms, and believe that assumptions ought to be both plausible and compatible with observations. This is probably the most urgently needed paradigm shift in economics.

NAEC seminar with Olivier Blanchard “Rethinking macroeconomic policy”, OECD, 5 July 2018 http://www.oecd.org/naec/events/rethinking-macroeconomic-policy.htm

The (unfortunate) complexity of the economy, Jean-Philippe Bouchaud, Physics World, April 2009, p.28-32 https://arxiv.org/abs/0904.0805v1

Steve Keen

One popular explanation of financial crises is excessive government debt and spending. Historical data going back two centuries reveal serious flaws in this argument, though. Government debt was falling before two major crises and started to rise after them, and even fell to zero before the 1837 crisis, and did not rise after it. Financial crises are caused by a boom-bust process driven by private credit: excessive private debt and credit before crisis, negative credit during it (the annual change in private debt being negative rather than positive). Private debt was rising before three great crises and started to fall after the crises began. There was negative credit in the 1837 crisis, as well as in 1929 and 2007.

Fluctuations in credit are of minor significance when private debt levels are low, but are catastrophic when they are high. The mechanism starts from the fact that expenditure is income. What you spend becomes income for someone else. There are two sources of expenditure: turnover of existing money; and new money created by exports exceeding imports, governments spending more than they tax, and banks lending more than they get back in repayments. Booms are caused by bank lending, as extra credit-money is spent into the economy. Bust follows when growth in private debt stops. Credit is not the largest component of expenditure but it is by far the most volatile. A fall in credit can cause crisis even if other factors are still growing, with the effect depending on both level and rate of change of private debt. We can see this by looking at a low-debt and a high-debt example.

In the low debt ratio example, imagine an economy with turnover of existing money of USD 1000 billion/year initially, growing at 10%/year. Private debt is initially 50% of turnover of existing money, USD 500 billion, growing at 20%/year. Credit is USD 100 billion/year. Total demand is USD 1100 billion/year. The next year, turnover of existing money is USD 1100 billion; growth of debt is 10%/year; credit = USD 60 billion/year (10% of USD 600 billion); total demand is USD 1160 billion/year: USD 60 billion higher than previous year.

For the high debt ratio example, we start with the same figures, except that private debt initially is not 50% but 200% of turnover of existing money, so USD 2000 billion, growing at 20%/year, meaning credit is USD 400 billion rather than USD 100 billion/year. Total demand is now USD 1400bn/year and turnover of existing money the next year is USD 1100 billion/year. Growth of debt slows to 10%/year. Credit is USD 240 billion/year (10% of USD 2400 billion). Total demand is higher than the low debt example, USD 1340 billion versus USD 1160/year, but this is USD 60 billion lower than the previous year in the high debt scenario.

Both the level of private debt/GDP ratio and rate of growth matter. The danger zone is when private debt is greater than 150% of GDP, and credit accounts for a large fraction of total demand (10% of GDP). The data on this for recent crises are overwhelming. Japan (the first country to suffer a serious credit crisis, back in 1990) the United States, United Kingdom, and Spain had crises when private debt reached historically unprecedented levels -150% in the United States, as high as 220% in Spain and Japan. All of these crises were preceded by credit rising to substantial levels, 20% of GDP in most cases, and almost 40% in the case of Spain (where credit is the annual change in private debt).

All involved negative credit, with Spain being the worst, with credit at negative 19% of GDP in 2013. This was unprecedented since the end of WWII for all of them except the United Kingdom, and even there, the previous negative credit events had been short-lived, whereas credit was negative for most of 2009-2015. Not only did these crises begin when the credit started to fall, the ups and downs of credit were a major determinant of economic activity at all times in every country except the United Kingdom since the 1990s.

Similar relationships exist in asset markets, and in particular, between mortgage credit and house prices. This contradicts the canons of conventional finance theory, which argues that leverage does not determine asset prices. The main determinant of the change in house prices is the change in mortgage credit. For US data, econometric testing confirms that changes in mortgage credit cause changes in house prices, rather than vice versa. The correlations between change in total household credit and change in house prices since 1970 for the other countries are respectively 0.4 for Japan, 0.6 for the United Kingdom, and 0.47 for Spain.

Data like these demand examination, but a decade after the financial crisis, mainstream economists continue to ignore them. These economists have learnt an intricate and superficially all-encompassing theory, which they believe provides not merely an explanation of the complicated reality of the economy, but also a guide as to how it can be improved. A core component of mainstream theory is the belief that one can quite literally ignore the banking system when modelling the macroeconomy. This was easy to do before 2007 since there had not been a banking crisis of this scale since the Great Depression. Voluntary blindness about the role of banks and credit in macroeconomics will give the world no warning again when another crisis approaches, not because no warning is possible, but because this wilful ignorance turns a blind eye to the very obvious causes of financial crises.

These causes are unfolding now, not at a global level but in the many countries that avoided the crisis in 2008 by continuing to accumulate more private debt. The four largest such economies are China, Canada, South Korea and Australia. Others include Singapore, Sweden, Norway and Belgium. In all of these countries, government policies led the private sector to avoid the negative credit experiences that made the 2008 crisis so severe in the United States and United Kingdom. But they accumulated even more private debt than the United States and United Kingdom in 2008. These countries will therefore experience their own, localised versions of the 2008 crisis when their credit bubbles burst. When their crises occur, virtually all the world’s major economies will be caught in private debt traps. The only exceptions will be countries like Germany that have exploited huge trade surpluses to enable their private sector debt levels to fall over the last two decades.

NAEC seminar “Can we avoid another financial crisis?”, OECD, 5 October 2017 http://www.oecd.org/naec/events/can-we-avoid-another-financial-crisis.htm

“Hiding in plain view: why economists can’t see the obvious coming” https://www.finance-watch.org/hiding-in-plain-view-why-economists-cant-see-the-obvious-coming

Sheri M. Markose

The 2007 financial crisis exposed the shortcomings of monetary economics and the regulatory framework known as Basel II. While financial innovations were progressing at a rapid rate, there was a lack of urgency to develop modelling tools capable of mapping and studying the massive interrelationships in the financial system implied by the workings of new financial products. Regulators, and other actors, had to rely on approaches dating from the period of double-digit stagflation in the 1970s and early 1980s when inflationary overheating was the sign of growing monetary and economic instability. The epochal reduction in inflation starting from about 1994 gave a semblance of calm and led to complacency.

A lack of a holistic perspective on the linkages between constituent elements can be blamed for why Basel II regulatory authorities encouraged bank behaviour that may appear sound at an individual level but contributes to system-wide failure. Systemic risk in financial systems, like environmental externalities which lead to overuse and degradation of resources, arises from design problems that are required to attenuate individual behaviour based on local incentives to prevent system collapse.

ICT based multi-agent financial network models can be useful in monitoring and analysing existing systems and can be used as computational test beds for the design of robust policy reforms. They can compensate for the weaknesses of mainstream macroeconomic or monetary models for policy that show an absence of the endemic arms race of strategic gaming by those regulated, and the weaknesses of econometric models that cannot handle structural interconnections and interactions between economic units.

Network models are increasingly being used to obtain a better understanding of stability of systems in biology, eco-systems, road transport, infrastructure and cities, engineering, power networks, information systems, etc. Network analysis and fine-grained firm level data based multi-agent simulators can also help address stability concerns for any financial market. Typically, in a financial network, the nodes are financial institutions and there are links called in-degrees which represent obligations from others, while out-degrees represent a financial entity’s obligations to others.

Network models depict causal chains between nodes rather than relying solely on statistical correlations which still remain the basis of most contagion models. The study of causal chains of network interconnections with nodes taken to be ‘agents’ with capacity for rule-based behaviour or fully autonomous behaviour that represents financial intermediaries (FIs) and regulatory authorities, constitutes the framework of financial network modelling.

The contractual obligations between FIs, and FIs and end users that determine bilateral flows of payoffs, constitute pre-existing network structures. A crisis with default of counterparties can trigger further contingent claims and large losses at default due to collapse in asset markets. Interactions of agents produce system-wide feedback loops. In the traditional equation-oriented analyses, structural changes from strategic behaviour and tracing of causal links and influences of feedback loops on individual decisions are almost impossible to do. In agent-based models, these need not be restricted to pre-specified equations that have to be estimated using past data in econometric or time series approaches. Agent-based ICT technology embedded in fine-grained digital maps of the structural interconnections of financial markets should therefore be developed as the starting point of stress tests and scenario analysis, especially in the context of the policy design.

Financial networks are not random and are most likely to have network properties like other socio-economic, communication and information networks. These manifest a statistical signature of complex systems, namely, a top tier multi-hub of few agents who are highly connected among themselves and to other nodes that show few if any connections to others in the periphery. The consequence of the clustered structure of a network is short path lengths between a node and any other node in the system. This is efficient in terms of liquidity and informational flows in good times, but worsens fragility in bad times when so-called hub banks (‘super-spreaders’) fail or suffer illiquidity. Failure of a big unit increases the probability of failure of other big units, an aspect of the too-interconnected-to-fail phenomenon. Structurally, however, the interconnected hubs can contain the liquidity shocks and prevent them from going to the extremities, but only if there are adequate buffers.

The presence of highly-connected and contagion-causing players typical of a complex system network perspective is to be contrasted with what economists regard to be an equilibrium network. In the latter, the probability that a contagion occurs conditional on one bank failing is significantly reduced, but the drivers of network formation in the real world are different from those assumed in economic equilibrium models. In terms of propagation of failure, however, it is not true that financial systems where no node is too interconnected are necessarily easier to manage in terms of structural coherence and stability. Stability analysis shows that the less-interconnected system is in some respects more dangerous. This suggests the need for caution in espousing an ideal network topology for financial networks.

It is important to consider network formation to be a complex adaptive process in that nodes interact strategically and respond to institutional incentives. A key aspect of complex adaptive systems is the capacity of interacting agents to show über intelligence with strong proclivities for contrarian (rule breaking) behaviour and the production of structure changing novelty and ‘surprises’. This takes the co-evolutionary form of a regulator-regulatee arms race with monitoring and production of countervailing new measures by the authorities in response to regulatee deviations from rules due to perverse incentives or loopholes. Failure to monitor and co-evolve the regulatory framework by authorities could result in system collapse.

Instability of large networks can result from a combination of individually rational behaviour and policy incentives which reinforce local efficiency but cause an increase in concentration and interconnectedness in the form of closer coupling with reduced buffers of nodes to a point of supercriticality or instability. The pressure to conserve scarce resources can lead to buffers being treated as costly and superfluous, leading to tighter coupling within the system. Economic forces can drive both designed and self-organising systems towards being balanced on the point of supercriticality where extreme system failure can follow. In the financial system, the different ways by which FIs in the system implement avoidance or reduction of key buffers (capital, collateral and margin requirements, for example) plus the numbers of those doing this have implications for the size of the hub nodes, the inter-connectivity between them and smaller nodes, and also contingent feedback loops of the system. All these factors can move the system to a supercritical state.

Socio-economic system failures, including financial crises, arise from a disparity between the pursuit of local interest and those needed for overall stability of the system. Poor rules made with no cognizance of their systemic risk consequences can wreck financial superstructures faster than any terrorist malfeasance.

NAEC seminar “Systemic Risk Analysis in Finance: New Approaches and Tools”, OECD, 9 September 2013

“Multi-Agent Financial Network Analyses For Systemic Risk Management Post-2007 Financial Crisis: A New Complexity Perspective for G10 and BRICs”, Sheri Markose, et al., Research Gate, 2010 https://www.researchgate.net/publication/267766719

Maurice Obstfeld

The crisis taught us three lessons, but these were there to be learned even before 2008: finance is central to macroeconomic outcomes; multiple equilibria can be all-important under stressful conditions; and the political economy of policy matters. The processes that would precipitate the fall of Lehman Brothers and provoke the crisis were already at work decades before in Asia and Latin America. There were even warnings about the rise of anti-globalist sentiment if those who lost out only had nationalism to turn to.

The crisis also highlighted three trilemmas, one monetary, one financial and one political. The classic monetary “trilemma” (a word coined by Friedman) postulates that countries face a trade-off among the objectives of exchange rate stability, free capital mobility, and independent monetary policy. If a country chooses exchange rate stability and free capital mobility, it must give up monetary policy autonomy; conversely, an independent monetary policy in the presence of free capital flows is possible through exchange rate flexibility. The rise in cross-border capital flows over the past few decades, and the frequent boom-bust cycles in capital flows, have however put the trilemma to the test and doubts have been raised about the ability of countries with flexible exchange rates to insulate their financial conditions from changes in key-currency financial centres.

In an alternative view, cross-border financial spillovers are similar for fixed and flexible exchange rate countries, implying the irrelevance of the exchange rate regime, and a two-way trade-off between capital mobility and monetary autonomy (Rey’s policy “dilemma” rather than a trilemma). According to this argument, regardless of the currency regime, monetary autonomy cannot insulate countries, so they need to use macroprudential policies, or failing that, capital controls.

There is however strong evidence from the response of a range of domestic financial variables to global financial conditions across exchange rate regimes in 43 emerging market economies that floating rates do provide a degree of insulation even from foreign financial shocks. In particular, exchange rate flexibility successfully dampens the magnitude of the cross-border transmission to domestic credit growth, real estate prices, and financial sector leverage. Global investor risk aversion shocks are transmitted more strongly through cross-border flows when the recipient countries have relatively inflexible exchange rate regimes and most emerging economies that have chosen a resolution of the monetary trilemma based on exchange rate flexibility have gained.

However, the bigger problem is the enhanced difficulty of effective financial policy in an open economy: the Schoenmaker or financial trilemma: financial stability, financial integration and national financial policies are incompatible at the same time. For example, strict rules on subprime lending have little impact if foreign banks still lend to these individuals. One response has been greater use of macroprudential policy, but this is not the whole answer and there are moves to improve international regulation as well as more questioning of the benefits from international capital movements and more openness to thinking about measures that target capital flows.

These flows are an integral component of multiple equilibria. While the classic literature focused on bank runs or currency crises, the 1980s Asian crisis highlighted “twin” crises, banking plus currency. But multiple equilibria are not just an emerging market pathology. The 2008 crisis showed the importance not only of bank runs, but of the flight of short-term wholesale funding from non-banks. The euro area crisis showed how sovereign debt could also be subject to multiple equilibria and the importance of feedback loops, non-linearity and other complex phenomena. Financial stability requires a fiscally strong sovereign, and sovereign weakness can undermine faith in financial safety net. Banking or financial system weakness may lead to government support, but if markets feel that government finances are too weak to carry the extra burden, interest rates will go up, making the burden even heavier and leading to sovereign debt worries. The sudden shifts this provokes between equilibria is a highly non-linear effect that cannot be accommodated in linear macroeconomic models.

This “doom loop” between fiscal weakness and financial fragility is illustrated by three problems: fiscal weakness worsens slumps because the government interventions lack sufficient means; the slump makes the fiscal situation worse and resources used for any interventions further reduce fiscal space to react to future crises; and recessions have hysteresis effects – the impact of the crisis on productivity lasts many years after the crisis ends.

The crisis also has political impacts, influenced by the fact that median real incomes have grown slowly or stagnated and inequality of income and wealth has increased in many countries (although these trends pre-date the crisis). While there is discussion of whether job dislocation; de-industrialisation; and the splits between high and low skill workers, urban and rural areas, and mobile and immobile workers are due to primarily to globalisation, technology, or policies, it is clear that these developments have fuelled resentment against traditional elites and “experts”, and the continued political power of financial interests, that finds expression in various populist and nationalistic movements.

This resentment can be translated into hostility against emerging and developing country economies, whose growth is outstripping that of the advanced economies. This is one factor making global co-operation harder and focussing attention on the Rodrik trilemma which says that at most two out of the following three are compatible: democracy; national policy autonomy; and extensive globalisation. The “sweet spot” where all three can coexist requires an inclusive policy framework such that most people gain from globalisation, even if inequality persists (as has been the case in many emerging and developing country economies). But designing and implementing such a framework is endogenous to voter choices, and there is no guarantee they would choose it.

So somewhat paradoxically, by helping emerging economies to succeed and thereby reducing the relative importance of the advanced economies, globalisation has threatened its own sustainability, and the future of multilateral co-operation. The challenge for economists is to find the best policies in the face of past and future structural transformation and to convince the public and politicians to adopt them. Scientific rigor remains necessary, but it is no longer sufficient. In 1924, another turbulent period, Keynes put it this way: the economist must be “as aloof and incorruptible as an artist, yet sometimes as near to earth as a politician”.

“10 Years after the failure of Lehman Brothers: What have we learned?” NAEC Conference, OECD, 13-14 September 2018, http://www.oecd.org/naec/10-years-after-the-crisis/

“Trilemmas and Tradeoffs: Living with Financial Globalization” BIS Working papers N°480, January 2015 https://www.bis.org/publ/work480.pdf

Robert Shiller

Narrative economics is based on the premise that narratives are not just a way describing and seeking to understand what has happened, but that stories that “go viral” evolve to actually affect outcomes, including crises, depressions, and recessions. The narrative basis of economic phenomena might be hard to see since narratives are not easy to measure, but by incorporating an understanding of popular narratives into their explanations, economists will become more sensitive to such influences and may produce better forecasts. For the last half century, one-year forecasts have been worthless on the whole.

Two elements are particularly important to narrative economics: word-of-mouth contagion of ideas in the form of stories; and people’s efforts to generate new contagious stories or make existing stories more contagious. An economic narrative reminds people of facts they may have forgotten, explains how things work in the economy, and affects how people think about the justification or purpose of economic actions. Seven propositions are key to economic narratives.

  1. 1. The timetable and magnitude of economic narrative epidemics can vary widely.

  2. 2. Narratives may be rarely heard and still economically important. People may not talk much about important the important economic decisions that are given a lot of attention in the media, but they will discuss the effects and fears they associate with the economy. During the Great Depression, for example, stories of hard-working people having to eat from garbage cans were contagious.

  3. 3. Narrative constellations have more impact than any one narrative, for instance stories around cryptocurrency featuring people who made fortunes from Bitcoin or would have made fortunes if they’d held on to them.

  4. 4. The economic impact of narratives may change through time, and we must resist the temptation to assume that all the narratives featuring the same words mean the same years apart.

  5. 5. Truth is not enough to stop false narratives.

  6. 6. Reinforcement matters, and the contagion of economic narratives builds on opportunities for repetition.

  7. 7. Economic narratives thrive on human interest, identity and patriotism.

The first narrative of the Great Depression was that of the stock market drop on October 28, 1929. This narrative was especially powerful, in its suddenness and severity, focusing public attention on a crash as never before in America. But, beyond the record size, it is hard to say what made this crash narrative such a success that it persists today. Part of the strength seems to come from a certain moralising. Sermons preached on the Sunday after the crash attributed it to moral and spiritual excesses, helping frame a narrative of a sort of day-of-judgment on the “Roaring Twenties.”

Another narrative at the beginning of the Great Depression was that of a repeat of the 1920-21 crash. For the general public, this would be falling prices, so it made sense to delay purchases. Economists expected the contraction to be as short lived as in 1920-21, which helps explain why President Hoover and others confidently explained that it would be over soon. But the public didn’t necessarily believe the President. Economists should look more at testimonies of women to understand the consumption function, and how they decided on shopping strategies, particularly since shopping was mainly a woman’s role.

Even as it happened, the contraction was thought of popularly as the product of some kind of feedback, famously tackled by President Roosevelt stating “the only thing to fear is fear itself”. Roosevelt took the unusual step of addressing the nation on the radio at a time of a massive national bank run that had necessitated shutting all the banks. In this “fireside chat” he explained the banking crisis and asked people not to continue their demands on banks. His personal request ended the run and money flowed into, not out of, the banks when they reopened. The narrative of this first fireside chat is still with us today, but the narrative has not been powerful enough, or not used well enough, to prevent recessions.

The macro storyline in the Great Depression gradually morphed into a national revulsion against the excesses of the Roaring Twenties. Contagion rates for stories of business failures, rather than inspirational stories, were naturally high at a time of high unemployment. Other narratives focused on the rising leftist or communist movement. The increasing radicalisation of President Roosevelt plays a part in these stories: in 1936, speaking of the magnates of organised money, he said “I welcome their hatred”.

Financial crises like 2008 are also driven by stories. Stories about bank runs in the 19th century were virtually synonymous with financial crises. After the Great Depression bank runs were thought to be cured, but the run on Northern Rock in 2007, the first UK bank run since 1866, brought back the old narratives of panicked depositors forming angry crowds outside closed banks. The story led to a nervousness internationally, and in 2008 to the Washington Mutual bank run in the United States, and the Reserve Prime Fund run a few days after that. These events then led to the very unconventional US government guarantee of all US money market funds for a year. Governments were aware that they could not let the old story of a bank run go live over concern for its effects on public anxiety.

Naming the financial crisis after the Great Depression was not the choice of any one individual. There had been earlier unsuccessful attempts to attach the name “Great Recession” to preceding recessions, so what is it about the 2007-9 event that made the name “Great Recession” suddenly contagious? Judging from the peak US unemployment rate, it was less severe than the 1981-82 recession. Perhaps it was because the 2007-9 event fitted the most generic and ill-informed memories of the Great Depression. People remember massive bank failures as part of the Great Depression story, and that was a better fit with the events of 2007. In the 1981-82 recession the stock market had not been booming, and the stock market did not fall below its 1980 value by 1982. In contrast, 2007-9 saw a near halving of the market from very high levels. People in 1981-82 were preoccupied with out-of-control consumer-price inflation, and saw the events in terms of a suddenly strong central bank effort to contain the inflation.

The Great Depression is exaggerated in people’s minds because of its legendary status. In 2007-09 presidents and prime ministers invoked parallels to the Great Depression to justify their requests to apply stimulus. Did this contribute to a self-fulfilling prophecy, a mirror event to the Great Depression, albeit not as severe (Great Recession, not Great Depression)? However, no politician can actually control the progression of the narratives they create. The manner in which these narratives unfold will play an important role in any economic forecast. To best predict economic activity, we need, among other things, to watch the narratives and ask: how will the emerging twists in the narratives affect propensity to spend, to start unconventional new businesses, to hire new employees? In short, how will the animal spirits be affected?

NAEC seminar “Narrative economics”, OECD, 10 September 2019, http://www.oecd.org/naec/events/narrative-economics.htm

"Narrative Economics," Robert J. Shiller, American Economic Review, vol 107(4), 2017, pages 967-1004

Narrative Economics: How Stories Go Viral and Drive Major Economic Events, Robert J. Shiller, Princeton University Press, 2019

William White

“It is worse than a crime. It is a mistake.”

Joseph Fouche

“We have gotten into a terrible muddle. We have blundered in the operations of a delicate machine, the workings of which we do not understand.”

John Maynard Keynes

The grave problems now facing the global economy have been building up over many decades. In large part they are the by-product of well intentioned but mistaken macroeconomic policies, themselves based on mistaken assumptions about the nature of the global economy. A philosopher would say policymakers have made a fundamental ontological error. Changing those underlying beliefs, and thus avoiding the disastrous effects of “still more of the same” macroeconomic policies, is the biggest challenge now faced by policymakers.

The analytical framework used by policymakers assumes that the economy can be adequately represented by linear models that are essentially simple and static. Moreover, the economy is assumed to tend rather mechanistically towards an “equilibrium” which has properties the policymakers want - like full employment. The economy is therefore understandable and easily controllable. Really bad outcomes are ruled out by assumption.

Unfortunately, this framework is fundamentally mistaken, as events over the last decade have shown. The economy is actually a complex, adaptive system (CAS) which is constantly evolving, never in equilibrium and has highly non-linear properties. Fortunately, CAS are ubiquitous in both nature and society, and how best to manage them has been well been studied by many disciplines. Still more fortunately, such systems share many basic characteristics. This communality implies that insights from other disciplines might be speedily applied to macroeconomic policies as well. Ironically, the conceptual embrace of complexity leads to at least ten lessons for policymakers that are actually quite simple.

Lesson 1: Policymakers’ multiple objectives make trade-offs inevitable. CAS break down on a regular basis, determined by Power Laws, so policy must always trade off efficiency against sustainability and system resilience. Macroeconomic polices must also take note of their implications for the distribution of income and wealth. Such considerations affect the transmission mechanism of macro policies and have important spillover effects into even more complex social and political systems.

Lesson 2: Policymakers can affect structure, and structure matters. While a CAS will have its own evolutionary dynamic, policy induced structural changes can make it easier to achieve desired objectives. It is well known that structural reforms can increase static efficiency. It seems less well known that buffers, redundancy and modularity can increase resilience. Unnecessary complexity should be stripped away.

Lesson 3: Policymakers should minimax not maximise. Our understanding of CAS will always be incomplete. Thus, optimisation is beyond our powers. Rather, since systemic breakdowns can have extremely bad outcomes, policy should focus more heavily on trying to avoid such outcomes. The use of highly experimental policies (and new products) should be constrained by the “do no harm” principle which guides both medical doctors and drug administrations.

Lesson 4: Policymakers should act more symmetrically. CAS are all path dependent and therefore “where you are” conditions “where you can go”. Stocks of debt built up over time make the economy increasingly fragile in all states of nature, both inflationary and deflationary. To avoid such buildups both monetary and fiscal policies should lean against upturns as strongly as they lean against downturns. Private sector debt exposures should be further limited by revoking interest deductibility for tax purposes as well as share buybacks.

Lesson 5: Policymakers should expect the unexpected. Because CAS are “adaptive”, policymakers are always in danger of fighting the last war. Worse, their own policies often encourage changes (regulatory evasion and moral hazard) that lead to this outcome. “Whack a mole” is not a strategy.

Lesson 6: Policymakers should focus on systemic risks more than “triggers”. In a highly stressed CAS, almost anything could be the trigger for a crisis. It is far more important to develop indicators of growing risks to systemic stability. That said, new financial products often “trigger” broader problems.

Lesson 7: Policymakers should be guided by multiple indicators. In a CAS, many things can go wrong. The belief that Consumer Price Index (CPI) “price stability” is sufficient to ensure macroeconomic stability is a false belief. Similarly, “financial stability” (the stability of the financial sector) is also insufficient since problems (like rising corporate debt or household debt) can arise outside of the financial system.

Lesson 8: Policymakers cannot forecast. In CAS, the future is essentially unknowable. While such systems can remain stable for long periods, a prediction that they will continue to do so is simply unwarranted extrapolation. Instead of decimal point forecasts, it would be better to provide alternative scenarios based on an assessment of emerging threats to systemic stability. This would also serve to remind people that radical uncertainty is a central characteristic of CAS. This implies the need for much larger buffers than are implied by traditional risk assessment procedures.

Lesson 9: Policymakers should be prepared for breakdowns. Since crises are inevitable in CAS, policymakers should prepare beforehand. “War games” should be played regularly, recognising that crises can start and unfold in myriad ways. Memoranda of Understanding between different policymakers should be negotiated and agreed. Legislation to ensure orderly insolvencies (for corporates, households and financial institutions) needs to be enacted. Since crises can vary in myriad ways, it is important that the authorities have the flexibility required to respond adequately.

Lesson 10: No policymaker is an island. By definition, CAS are defined by their interdependencies. All national policymakers must therefore formulate their policies with a view to the effects on other national policymakers and their responses. Such considerations evidently limit the “independence“ of central banks and of national regulatory agencies. Moreover, since the interdependencies are increasingly international, this raises questions about the viability of policies directed purely to national objectives. Embracing complexity requires a review of the existing International Monetary (Non) System.

Economists are fond of saying “It takes a model to replace a model”. We now have such a model. The economy is a complex, adaptive system and should be treated as such. The ten practical lessons for policymakers, described above, are simple but revolutionary. That is precisely why they should be adopted. We need a paradigm shift to get off our current disastrous path. If we continue to pursue relentlessly the impeccable logic of an argument which is based on false assumptions, it will be worse than a crime. It will be a mistake.

Matheus Grasselli

Looking back 10 years after the last crisis, several participants in the NAEC conference that took place in September 2018, myself included, expressed concern that, when the next crisis inevitably hit, central banks would find themselves much more constrained in their response, for both regulatory and political reasons. Covid-19 showed that these concerns were spectacularly misplaced: central banks around the world did more in a few weeks to contain the fallout from the pandemic than they had done in several months following the 2008 crisis and its aftermath.

Key measures of financial distress, began to increase in late February 2020 and headed to crisis territory in early March, when it became clear that the pandemic would cause widespread disruptions in the world economy. Stock markets experienced gyrations of such rapidity that even automatic circuit breakers were not fast enough to halt losses. The financial press ran out of nicknames for crash days: Black Monday I (March 9) was followed by Black Thursday (March 12) and then Black Monday II (March 16) – each breaking records set in previous crises.

As governments were enacting measures to contain a rapidly spreading virus, central banks sprang into action to prevent the spread of an even faster contagion in financial markets. In what follows, I focus on the measures taken by the Federal Reserve, but similar stories can be told about the Bank of England, the ECB, and other major central banks.

The first line of defence of central banks in a crisis is to accelerate conventional monetary policy. And accelerate the Fed did: in two consecutive rate cuts in as many weeks, it brought the policy rate down by 150 basis points to its lowest possible corridor above zero. By comparison, it took the Fed 9 months and 4 rate cuts between March and December 2008 to lower the policy rate from 2.25% to effectively zero.

The next line of defence consists of a central bank assuming the role of lender-of-last-resort (LOLR) for solvent banks and financial institutions facing immediate funding liquidity shortages. The traditional “discount window” falls into this category and has been renamed by the Fed as Primary Credit, Secondary Credit, and Seasonal Credit, depending on the institutions that have access to it and the rates charged. The use of these facilities spiked immediately after the Lehman weekend, rising at a pace of approximately USD 100 billion per week and peaking at about USD 450 billion on October 15, 2008. This was the same pace of expansion observed in mid-March 2020; however, the discount window was further extended by the creation of the Term Asset-Backed Securities Loan Facility offering non-recourse loans to issuers of asset-backed securities (with a non-resource loan, the lender cannot claim the assets of the borrower other than those used as collateral). In so doing, the Fed extended its role as LOLR to traditional depository institutions to include a much larger array of eligible borrowers than in 2008.

An essential feature of global markets highlighted by the 2008 crisis was that, because foreign banks have disproportionate quantities of dollar-denominated liabilities, global funding liquidity shortages are essentially shortages of US dollars. To address this, the Fed made extensive use of swap lines with a select group of central banks in 2008, peaking at slightly less than USD 600 billion. These swap lines have remained in place since then, reaching a level of about USD 100 billion during the Euro crisis in early 2012, though they were rarely used in following years. One of the key question marks about international financial co-operation in the Trump era has been the extent to which swap lines would be used again in the next crisis. The response to Covid-19 put this question to rest: swap lines were extended to a total of 14 central banks, including those of Brazil, Mexico, and South Korea, and quickly reached their legal maximum of USD 450 billion, where they currently sit.

A novel aspect of central bank response in 2008 was the role of dealer-of-last-resort (DOLR). This was a consequence of the integration of modern banking and finance into what Perry Mehrling describes as “money market funding of capital market lending”. In other words, instead of intermediating between traditional loans and deposits, banks and other financial institutions are now part of a complex network of interconnected balance sheets of market-makers straddling securities deals all the way from primary savers to ultimate borrowers. Such dealers hold inventory risk and make profits on the basis of bid-ask spreads in return. The problem arises when, in a crisis, no spread is large enough for private dealers to “make” markets, which simply cease to function. By absorbing a vast array of securities in its own balance sheet, a central bank can restore the ability of these markets to operate. This type of intervention to provide market liquidity, as opposed to more traditional funding liquidity, was behind the most audacious moves made by the Fed in 2008, quite distinct from the more common rationales for quantitative easing related to interest rates. The full extent of these interventions is difficult to measure, as many were done through special purpose vehicles (SPV) whose assets do not appear on the Fed’s balance sheet, but data for one of them, the Commercial Paper Funding Facility, shows interventions of the order of USD 250 billion at its maximum. This DOLR role for the Fed has made a comeback during Covid-19, with the re-opening of several 2008-era facilities and the creation of brand-new ones. For example, as part of what is being called Jeremy Powell’s “whatever it takes moment”, on March 23, the Fed announced a new facility, established as an SPV with equity provided by the Treasury as part of the CARES Act, with the purpose of directly purchasing corporate bonds – a move previously considered unthinkable for a central bank.

The combined results of all these measures is that, between March 6 and May 27, 2020, the Fed’s balance sheet expanded by a factor of two thirds to a total size of over USD 7 trillion – an approximate pace of USD 220 billion per week, more than double the monthly average expansion observed in the aftermath of the 2008 crisis. In both scale and scope, then, the response to Covid-19 appears to have inaugurated a new era in modern central banking.

Avner Offer

Between the early 1980s and 2008 domestic credit in the United Kingdom rose about fivefold as a percentage of GDP and about half of that in the United States. The mechanics of banking indicate how this expansion might have been achieved, and who stood to gain or lose.

Contrary to intuition, the constraint on lending is not prior funding, but the need for credit-worthy borrowers. To see how, begin with a pure credit economy with a single bank. The bank lends by crediting a borrower’s account, ‘by the stroke of a pen’, no more. When the borrower spends, the money never leaves the bank, it just moves to another account. Money is created when the bank lends, and is destroyed when repaid. The bank stands ready to lend to any reliable borrower. It has to be prudent since bad debts diminish capital and can threaten solvency. Profit arises from the real-world productivity of investment (the increment of money comes out of loans to other borrowers). If we started a banking system from scratch, this might be the model to use; if the single bank was also a central bank it could pursue macroeconomic objectives.

A bank can run out of money when there is more than one of them. The money lent out by one bank is likely to be deposited in another. For the first bank this is offset by inflows originating in loans made by its peers. Banks can create new money by lending and if they all do so at the same pace, most of it comes back with little need for extra funding. But they don’t. They compete for market share. Money goes out and comes in unevenly (with leakages to cash and overseas). Outgoings and incomings are settled in central bank reserves, which banks cannot create. Banks acquire these reserves by purchasing government bonds with their own deposit money. They sell or lend the bonds to the central bank in exchange for the reserves they need.

A commercial bank must be ready to pay out at all times but a good deal of its outgoings are not under its control. Credit lines, overdrafts and deposits can be drawn down at any time so the bank needs to be sure of access to central bank reserves. The bank’s own reserves there are only a thin buffer. The first line of defence are credit lines with other banks and the bank stands ready to lend any of its own surplus. If a bank cannot clear its balance, the central bank will do this automatically for a price. In the past banks used to hold a buffer of liquid assets such as commercial and treasury bills, but since the 1980s liquidity is mostly managed by borrowing short-term in the money markets. UK Banks used to keep asset maturities short but since the 1980s they are increasingly locked into more lucrative but long-maturity mortgages (American banks securitise their mortgages and sell them off). Liquidity risk and the need for funding arises from this maturity mismatch. Banks are meant to be in surplus overall but daily balance sheets can be volatile and they generate ‘liquidity gaps’. The central challenge of banking is to remain liquid at the lowest cost, which is as much an art as a science.

The cash floats of households and business (held as deposits) are the largest, most reliable and cheapest source of funding. Another source of liquidity is ‘shadow banking’, financial institutions without central bank accounts. These insurance companies, pension funds, hedge funds, fund managers and suchlike lend more in the aggregate than the banks. The financial sector as a whole borrows more than households, non-financial corporations, and government (each separately), most of it for unproductive speculation in securities.

After the banking deregulation of the early 1980s aggregate domestic debt service has risen to an order of 10-20 percent of GDP a year, a transfer from spenders to hoarders that is more than the cost of the health service in the United Kingdom. Deregulation in the 1980s facilitated credit expansion. Before the 1980s housing was mostly funded by building societies (in the United Kingdom) and to a lesser extent by ‘thrifts’ (Savings and Loans banks, in the United States). These institutions acted as intermediaries, taking cash from savers to lend out as mortgages without creating new money. House prices were kept in check by the limited supply of savings. In the 1980s commercial banks moved into housing finance but without the same prudential constraint. They could create new money. Housing is credit-worthy due to its built-in collateral, and abundant credit acts to raise its price. Shelter is an essential good, paid for out of household income, which constitutes a massive debt-service funding base. For borrowers, a rising proportion of income went into debt service, which was more than offset for them by rising house values. But debt service is self-limiting – eventually it diverts income away from consumption and employment, and instigates financial crises.

When the crises came governments and central banks took care of creditors ahead of any other group in society, with the largest outlays seen in peacetime. The banking system was forgiven its debts while governments imposed austerity on public services. Abundant credit has driven house prices beyond the reach of younger people and of those on moderate earnings. The suppression of demand by debt service, both public and private, has also held down household incomes. The consequences of these policy choices is a simmering dissatisfaction which has fed into a social and political crisis with no end in sight.

“The mystery of banking: an exploratory essay”, Avner Offer, unpublished paper, July 2020. https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxhdm9mZmVyfGd4OjU5ZDlhNzA2YjUwNWFjNTk

“The market turn: from social democracy to market liberalism”, Avner Offer, Economic History Review, vol 70 (4), 2017, pages 1051-1071

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2020

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.