The Romantic Economist addresses the limits of knowledge in markets
- Extra Reading
Richard Bronk, Visiting Fellow, European Institute, The London School of Economics and Political Science
The relationship between theory and fact is particularly problematic when interpreting the complex, creative and socially-constructed world of markets. This lecture will explore how theories frame our vision and how prices and institutions transmit information, before examining the vital distinction between measurable risk and radical uncertainty.
The Romantic Economist addresses the limits of knowledge in markets
Since the economic and banking crisis reached its peak eighteen months ago, there has been no shortage of explanations for the debacle. Most of these focus on the venality of some bankers, the perverse impact of remuneration structures on incentives, revealed failures in light-touch regulation, and the excessive build-up of private credit and global economic imbalances. All these factors have played their part. But I want to argue today that there were three profound intellectual errors underlying the construction and use of the dominant models of economics and risk analysis that may have played an equally important role. These errors concern the nature and limits of knowledge in markets and the vexed relationship between theories and facts. In a nutshell, I will argue that, before the crisis, we knew a lot less than we thought we knew: we exaggerated the knowledge conveyed by market prices; we gravely overestimated our ability to measure and predict future risks and, by misunderstanding its link to innovation, we underestimated the ubiquity of unquantifiable uncertainty. At the same time, we became fatally reliant on what we believed to be our best models without realising how far they structured and constrained our vision.
Three years ago, large swathes of the economics profession and of Wall Street and the City of London bought into some variant of the efficient markets hypothesis. This posits that market prices are reliable indicators of fundamental value - because market participants are forced by competitive pressures to make optimal use of available information, to avoid systematic errors in their forecasting, and to update their expectations rationally in the light of new evidence. Any deviation of market valuations from underlying fundamentals should be essentially random and short-lived. This hypothesis had enormous implications: markets were self-adjusting and market prices were the best signals of emerging fundamentals and underlying value. Government and regulatory intervention might be needed to deal with information asymmetries and improve the transparency of information between market participants; but, beyond this, second-guessing the wisdom of markets was unlikely to be a good idea. Suffice it to say, this cheerful and optimistic creed that the market knows best and efficiently prices available information and knowable risks has taken a knock since 2007. Market expectations and pricing are revealed to have been close to delusional for a long time and driven by successive waves of exuberance and despair - by Keynes's 'animal spirits'. So where did the efficient markets hypothesis go wrong?
From the mid-1990s onwards, financial markets, regulators and much of the rest of the corporate and public sectors fell under another spell - what my LSE colleague, Michael Power, has called the 'grand narrative of risk management'.[i] A whole range of new Value at Risk (VaR) and other models promised to calculate the risk of future loss or default on the basis of extensive analysis of data on the past; while an associated institutional culture of control systems and risk officers sought to manage decisions taken in uncertainty in a rational and scientific manner. As Power argues, the result of this seemingly defensive narrative was a dangerous illusion of control. The related illusion was, to quote from the Bank of England's Andrew Haldane, that we had entered a 'new era' of 'simultaneously higher return and lower risk', resulting from 'a shift in the technological frontier of risk management'.[ii] In fact, as I shall argue, the narrative of risk management depended in large part on a failure to understand the vital distinction between measurable risk and true 'unmeasurable' uncertainty. Fifteen years after Frank Knight first articulated this distinction,[iii] John Maynard Keynes famously said in The General Theory: 'The outstanding fact is the extreme precariousness of the basis of knowledge on which our estimates of prospective yield have to be made. ... [Our] existing knowledge does not provide a sufficient basis for a calculated mathematical expectation'.[iv] But Knight and Keynes were largely ignored in the heady days of the recent derivatives boom and rampant risk modelling. In the second half of the lecture, I will outline the nature and causes of this 'unmeasurable' uncertainty and assess to what extent it could ever be amenable to being turned into measurable risk. I will then show how this uncertainty affects the information conveyed by market prices and the nature of market expectations; and I will suggest some better approaches to decision-making in conditions of uncertainty.
But let me begin by tackling another intellectual error that I believe contributed to the crisis - the failure to understand how far the models and theories we use structure and constrain our vision. Only by understanding this can we explain how the increasing reliance on one set of economic theories and risk models, and the convergence on apparently 'best practice' in finance and regulation, left us so exposed to analytical myopia and practical disaster. I will argue that, while the dominant economic and risk models did focus attention on certain aspects of what was happening, the analytical monoculture they implied also blinded policy makers and bankers to emerging problems.
Now in my bid to address these three related intellectual errors, I am going to make use of the ideas of several historical figures within economics. But as the title of this lecture suggests, I will also look in a much more unusual place for some answers, namely the poetry and philosophy of the Romantics living about two centuries ago and their sustained critique of narrow rationalism and mechanical thinking. But why turn to the Romantics?
Let me answer this by saying how I first became interested in doing so. Before joining the London School of Economics in 2000, I worked for seventeen years in the City of London and, the longer I did so, the more aware I became of an important mismatch between the way economists usually model economies and the way markets often work in practice. In a nutshell, economists use equilibrium-based models and assumptions of rationally optimising individuals, while the markets I saw around me were dynamic and creative processes, driven by relentless innovation, self-reinforcing spasms of euphoria and despair, and the massive uncertainty these imply. There were some regularities of behaviour, of course, but as often as not these seemed to be a function of socially learned narratives and ideological frames of reference. And it suddenly struck me that, if economists wanted to make better sense of these phenomena, there was much they could learn from the Romantics, who had analysed in unsurpassed depth the role of language and metaphor in structuring beliefs, and the role of emotion and imagination in driving behaviour. The result, nearly a decade later was my book, The Romantic Economist,[v] copies of which just happen to be on display at the back of the hall.
Perhaps the most important lesson of Romanticism concerns the role played by imagination, metaphors and models in structuring our visions, beliefs and therefore social reality itself. The Romantics were clear that you never have access to a definitive and objective way of looking at the world; rather the world-as-it-appears-to-you is - at least to some extent - a creation of your mind. Your mind does not passively record and reflect facts 'out there'; nor do your beliefs merely imitate reality. Rather, if you are going to make sense of the chaos about you, you must supply an intellectual framework, a metaphorical colouring, a principle of selection. As Wordsworth put it, you half-create the world you see; your mind is a 'creator and receiver both'.[vi] Any particular observation you make is the joint product of sense data your mind receives and the conceptual or narrative structure your mind contributes to make sense of them. As Coleridge said, when arguing with a young scientist who thought he could analyse facts without first having a theory:
'You must have a lantern in your hand to give light, otherwise all the materials in the world are useless, for you cannot find them, and if you could, you could not arrange them'.[vii]
You cannot do without models and metaphors to help you understand the world, any more than you can do without a lantern to see in the dark. But the trouble with lanterns - and with theories and metaphors for that matter - is that the light they cast, the focus they bring, is limited and partial. This means that if you only use one light - one theoretical framework - you will keep stumbling on aspects of reality outside the area illuminated by your theory. To put it another way, the lens of metaphor or theoretical model can bias and distort your vision as well as focus it
It is exactly this that has been the recent fate of much of the economics, financial and policy-making fraternity. Before the economic crisis erupted, most central bankers, treasury officials and financiers were so convinced that the Greenspan approach to monetary policy, the efficient markets hypothesis and neo-classical economic models were sufficient, and had so internalised this one perspective, that they were simply not predisposed to see the asset price bubbles and other problems that were emerging because their theoretical and conceptual framework had no place for them. Similarly, bankers were so in the grip of Gaussian distribution-derived VaR models that most of them genuinely believed that what has since happened was unlikely to occur even once in the lifespan of the universe. It is not surprising that banks were not ready for the financial tsunami that hit them if they were relying on risk models that told them that the daily market moves we then saw in August 2007 were as much as 25-standard deviation events.[viii] Worse still, as Michael Power has pointed out, VaR models and related metrics of risk adjusted return on capital (RAROC) became more than a 'best practice' frame for the views of management about the risks they were running; they also became part of an 'increasing conceptual convergence between regulatory management of economic capital and internal business models.'[ix] This elision between the previously distinct perspectives and cognitive frames of regulator and regulated under the Basel-II regime was to prove disastrous. As it turned out, partially blind bankers and traders were being regulated by those with exactly the same type of myopia. The Romantic lesson to be drawn is that there is never only one right way of looking at a set of issues. Theoretical or modelling dogmatism in the name of best practice makes you like a horse wearing blinkers, good at focusing straight ahead on one thing, but liable to miss what is coming at you from left field. For this reason, it is crucial that we learn how to experiment with different metaphors, models and theoretical frameworks - different ways of seeing. Changing the metaphors we use helps us make sense of different aspects of reality and spot unexpected new developments.
In my book, The Romantic Economist, I argue that economists will have a better chance of making sense of the dynamic, social and creative aspects of markets if they experiment with organic metaphors from Romanticism instead of the equilibrium metaphors of 'social physics'. Coleridge pointed the way, in one of his lay sermons, when he attacked the notion that the economy is a 'self-regulating machine' that always reverts quickly and painlessly to equilibrium after a market crash. 'Persons are not things', he thundered, and 'man does not find his level'. The mechanical metaphor, he pointed out, overstates the likelihood in real life of ever reaching equilibrium and incorrectly treats people as inanimate commodities. For this reason, Coleridge advocated a more organic conception of state and economy. In a particularly prescient passage of the same sermon, he likened the spread of bankruptcies to 'a fever, at once contagious and epidemic';[x] and this organic simile, merely suggestive in his day, now finds echoes in recommendations by Andrew Haldane and others that we should learn from the field of epidemiology how to model and manage the dynamics of default risk and market panic.[xi] By borrowing models normally applied to 'flu or Sars epidemics, economists and financiers can isolate and simulate many of the threshold effects and self-reinforcing dynamics in markets that render them so unpredictable - dynamics not necessarily reflected in the historical correlations captured in standard risk models. As Haldane notes, these epidemiology modelling analogies can also suggest new approaches to regulation - such as focusing on the risk to the system posed by the most interconnected institutions, which in times of financial contagion act like 'super-spreaders' of disease. Other organic metaphors (such as the management of forest fires) suggest the importance of building 'firewalls' - of breaking up global market networks into smaller units so that the whole is less vulnerable.[xii]
The point of such experimentation with new organic models and perspectives is not to reject the old mechanical metaphors and models entirely or deny their analytical power; rather it is to understand why they sometimes focus attention effectively and other times fail lamentably to disclose what is important. My thesis is that standard equilibrium models in economics - with assumptions of rationally optimising individuals - and standard Gaussian risk models are much less successful when economists and financiers are modelling innovative markets, network interdependence and uncertainty. This leads me to suggest the need for clearer boundaries of applicability for standard models. This will bolster their effectiveness. It will also allow them to be supplemented with new models and metaphors better suited to working outside these boundaries.
It will seem strange to many economists and risk managers to advocate such modelling eclecticism. For much of the focus in both professions has been on finding a single unified theory and a coherent set of models that best represent and predict reality. But as Iris Murdoch wrote in her book on Jean-Paul Sartre's philosophy:
'What does exist is brute and nameless, it escapes from the scheme of relations in which we imagine it to be rigidly enclosed, it escapes from language and science, it is more than and other than our descriptions of it.'[xiii]
I would argue that we have no hope of capturing the multifaceted nature of reality unless we use a plurality of theoretical and modelling perspectives. Modern physicists, for example, no longer rely solely on the equilibrium models of Newtonian mechanics, however useful they are to build bridges. Moreover, there is to date no unified theoretical framework that can encompass the additional respective insights of relativity theory, quantum mechanics and chaos or complexity theory. And if there is no unified theory in physics, how much less likely there is to be one that works in economics or risk analysis. For not only does economics have to cater (as we shall see) for the 'Knightian' uncertainty caused by the human imagination; it also has to interpret a pre-interpreted world.
The behaviour studied by economists and risk modellers is already structured - in part at least - by the socially-formed narratives, conventions and economic theories that individual actors have internalised. This means that you cannot fully explain or predict economic and market behaviour unless you have learned to empathise with (the better to interpret) the various mind-sets and conceptual structures that influence beliefs and reasons for action. Without analytical imagination - the conscious effort and unconscious ability to place yourselves in the conceptual shoes of other market participants - you are always liable to miss key aspects of what is going on. The best investors in markets have long understood this. As Keynes pointed out, the key to successful investing, especially in the short-run, is to anticipate the shifting interpretations and conventional frames of other investors in the market.[xiv] And as George Soros has noted with his thesis of reflexivity, the bias and views of investors often affect the underlying fundamentals and influence the long-term future as well as prices in the short-run.[xv] New era stories affect the relative cost of capital across different sectors for years at a time. Narratives of bank failure (however initially unjustified) have a habit of becoming self-fulfilling if depositors react by withdrawing their savings. And a theoretical monoculture creates homogeneity of behaviour and high correlations in markets that can be truly terrifying. One of the many factors left out of risk models in the run-up to this crisis was the destabilising rise in correlations caused by the rapid internalisation of the same return-on-equity strategies, the same accounting conventions and the same risk models across so many markets, all in the name of best practice and regulatory harmonisation. Never again should economists, risk officers and regulators ignore the extent to which dominant narratives, theories and norms construct behavioural regularities. They need to become sociologists and anthropologists quite as much as economists and quantitative modellers and understand the impact of different narratives and perspectives.
Let me now turn to another intellectual error that helps explain the recent crisis - the failure to understand the impact of human creativity and innovation on the prevalence of uncertainty. Inadvertently echoing the Romantic essayist William Hazlitt, an economic philosopher writing some thirty years ago, George Shackle, emphasised the imaginative genesis of much of the uncertainty we face. He wrote of our 'own original, ungoverned novelties of imagination ... injecting, in some respect ex nihilo, the unforeknowable arrangement of elements'.[xvi] The future, that is, is unknowable because it is still to be created by the original choices we (and others) will make and the new possibilities we (and others) will imagine. For Shackle, 'Tomorrow is figment', and our economic expectations are both the creation of our imaginations and creative of the future. His message is corrosive of the standard notion that forward-looking market valuations can be stable and efficiently priced - that there is a static reality 'out there' on which rational expectations will converge thanks to competitive pressures. As Shackle put it, 'Valuation is expectation and expectation is imagination'.[xvii]
In practice, of course, the future is not a complete 'void' as Shackle seemed to suppose. It is in part rationally predictable - and some of the risks can be forecast - given observed and socially constructed regularities in behaviour. But the important point remains that Shackle was much more right than most modern economists and risk modellers would ever acknowledge, at least in situations where novelty and innovation abound. They have tended to ignore the fact that imagination and novelty create uncertainty, break the predictable links between the past and the future, and undercut the rationale for making probability forecasts on the basis of historical frequencies. A novel idea or innovation injects some entirely new elements into the equations of social life - think, for example of the Internet, the impact of which was almost entirely unpredictable ahead of its invention. Innovations often change the economic parameters and disturb previously stable regularities of behaviour. In other words, genuine uncertainty is an inevitable by-product of innovation and the capacity to imagine new options.
This is highly topical. For the assumption implicit in most risk models used by banks before the credit crunch was that uncertainty about the future could be turned into measurable risk on the basis of patterns and correlations trawled from data on the past. All you needed was a clever enough model and enough data. But regularities in past behaviour were very unlikely to be safe predictors of future risks when the banking industry itself was innovating furiously - bringing out complex new products each week that made it impossible for the future to resemble the old financial environment in which the data were gathered. In other words, risk modellers were ignoring the central distinction that Knight made between measurable risk and radical uncertainty.
Knight used the word 'risk' to designate 'measurable uncertainty': here the possible outcomes are known; they can be classified in groups and assigned probabilities or projected distributions 'either through calculation a priori or from statistics of past experience.' This is the realm of classic insurance markets - like fire insurance or life assurance - where the future can be assumed to be a shadow of the past. 'Uncertainty', by contrast, was the name Knight gave to cases where no probability can be computed because, for example, the case is unique.[xviii] This same distinction was in Keynes's mind in The General Theory when he wrote that 'human decisions affecting the future, whether personal or political or economic, cannot depend on strict mathematical expectation, since the basis for making such calculations does not exist.' For Keynes, uncertainty was magnified by market speculation, 'animal spirits' and 'waves of optimistic and pessimistic sentiment ... where no solid basis exists for a reasonable calculation'.[xix] But he failed to theorize extensively about the causes of uncertainty as opposed to its impact.
To get a feel for how negotiable this boundary between Knightian risk and uncertainty might be - for how far the risk management industry and regulators were justified over the last decade in implicitly or explicitly assuming they had turned uncertainty into measurable risk - it is essential to analyse more carefully the different kinds and sources of uncertainty in markets. To do this I want expand on two distinctions made by Robert Skidelsky in his most recent book on Keynes.
The first distinction is between 'asymmetric information' and 'symmetric ignorance'.[xx] Information asymmetries, where one party to a market transaction has an information advantage over another, have been extensively analysed in modern economic and regulatory theory. Such asymmetries can lead to opportunistic, even fraudulent, behaviour, to the mispricing of deals, or to thin markets characterised by such distrust between parties that trades dry up. The solution to such problems is at least theoretically straightforward: more transparency and disclosure of information should ensure that markets work more efficiently and that risks are priced correctly. The second sort of information problem is both more ignored and harder to solve, namely that of 'symmetric' ignorance - of genuine uncertainty faced by all parties. The key question then is how many of the information problems in this crisis were asymmetric and how many were symmetrical cases of real uncertainty. This is, of course, partly an empirical question that I cannot answer today. But a theoretical analysis of the causes of uncertainty can give us some clue.
This brings me to Skidelsky's second distinction - this time between 'epistemological' uncertainty, where relevant probabilities are unknown, and 'ontological' uncertainty, where they are logically unknowable.[xxi] Let me expand on this distinction. Epistemological uncertainty includes the difficulty of grasping all the multifaceted aspects of what is going on - particularly if locked into analysis with one set of models - and the sheer volume of information to be processed. It also relates to the difficulties of understanding the non-linear dynamics of complex systems and the self-reinforcing emotional dynamics of market confidence and panic that make prediction fraught. Some progress has been made, and is possible, in shifting the boundary between such epistemological uncertainty and measurable risk, even if non-linear and behavioural dynamics are not amenable to simple frequency distributions and precise definition of the likely spread of future returns. Ontological uncertainty, by contrast, implies the impossibility of knowing (even the categories and possible nature of) what has yet to be created or yet to evolve. This is the sort of uncertainty implied by radical innovation that revolutionises the parameters of markets and the range and nature of possible outcomes; and some non-linear dynamic systems may in time also allow for the emergence of genuine novelty. This kind of uncertainty can never be turned ex ante into measurable risk. The future opportunities and dangers we face are simply unknowable at the outset, and we must learn and adapt as we go along.
It is clear that in the last two decades both epistemological and ontological uncertainty have increased. Financial markets and products have become much more complex; and the volume of relevant information has swamped the mental capacity of any market participant to comprehend it, and arguably outstripped even the growth in computer processing power. Network dynamics in increasingly interconnected markets imply the need for less hubris about our capacity to predict exact risk profiles; while the importance of shifting narrative and emotional frames demand new more pluralistic modes of analysis that are rarely forthcoming. As for ontological uncertainty, we need look no further than the entirely novel dynamics in both primary and secondary markets caused by the invention of new products that securitised mortgages and bundled them up in ever more exotic ways. It is hard to see how historical data on default risk in the US housing market or in secondary markets could have been relevant after the scale of such innovation. And yet such data was used in many of the models assessing risks in the CDO and related markets.
In short, the risk models on which so much of the edifice of modern finance depended in the run-up to this crisis underestimated the epistemological uncertainty in modern markets and ignored the ontological uncertainty caused by rapid innovation. The complex non-linearities of financial networks and the multi-valence of social reality were rarely assessed and could not easily be codified in the dominant VaR models; while the ontological problem of innovation was assumed away in Gaussian risk models that assumed that you could read the standard deviation ranges of future outcomes from the distribution of past returns. In one revealing sense, though, the problem of innovation was not totally ignored. It was often argued that it was preferable to base risk calculations on high frequency data from the recent past, because too much had changed for data from thirty or more years ago to be relevant. But this argument was as inconsistent as it was dangerous. If only the recent past was considered relevant to the conditions of the day because of the parameter-altering nature of innovation, it should have been obvious that the recent past might not be relevant to the near future either. At the same time, the practice of collecting data only from the period now dubbed 'The Great Moderation' excluded readings from the market upheavals of the 1930s and 1970s that might at least have given a better idea of the scale of instability that can suddenly engulf dynamic markets operating in uncertainty.
One final point about Knight's argument before I leave the subject of risk: Being a good Chicago economist, Knight believed that that a competitive system will only allow profits to be made if there is genuine uncertainty:
'Profit arises out of the inherent, absolute unpredictability of things, out of the sheer brute fact that the results of human activity cannot be anticipated and then only in so far as even a probability calculation in regard to them is impossible and meaningless.'[xxii]
This is an interesting idea. In a competitive market, predictable profits are quickly competed away in the absence of monopoly rents or asymmetries of information. Now it is hard to argue that financial markets have recently been monopolistic. So were the giant profits in the sector a sign of the exploitation of information asymmetries? To some extent, they were. But they were also a result of the continual innovation of new products and the ability to exploit first-mover advantage; and such innovation came at the price of massive uncertainty. Regulators should perhaps have seen high profits in the banking sector as a sign of building uncertainty. The masters of the universe were making huge returns by playing with radical uncertainty.
So where does all this leave the efficient markets hypothesis with which we began? And what use can we make of market pricing to solve the problem of uncertainty? The answer I believe requires us to blend the insights of those two sparring partners in the economic study of the Great Depression - Hayek and Keynes. Hayek did not believe that the market ever reached an optimal equilibrium and would have been impatient with some of the glib claims of the efficient markets hypothesis. But he did believe that the price system in a free market is an unrivalled 'marvel' in its function as a coordination mechanism for communicating information. For Hayek it was precisely the role of market prices in helping each of us discover more about our uncertain predicament that made it so superior to socialist attempts to calculate our needs. As he wrote:
'...the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.'[xxiii]
Hayek's argument was that the signals given by market prices communicate summary information based on the innumerable decisions of individuals and so reflect key attributes of the dispersed and contextualised knowledge frequently available only to those individuals. For Hayek the market was a 'discovery procedure'[xxiv] for solving what I have dubbed epistemological and ontological uncertainty; and it was the very pluralism of individual perspectives that helped give the system epistemological robustness and importance.
Very few students of mine who grew up in Eastern Europe before 1989 and watched repeated failed attempts to find solutions to consumer and corporate needs without the help of a functioning price system have doubted that Hayek was here articulating a crucial truth about the value of markets. But it is only a partial truth. For sometimes, as we have seen in recent years, market prices can be profoundly misleading and their movements positively destabilising. To understand why we need to turn to Keynes. In chapter 12 of The General Theory, he argues that in conditions of uncertainty (where no basis exists for 'calculated mathematical expectation') investors frequently fall back on a conventional approach to valuations - in particular 'assuming that the existing state of affairs will continue indefinitely, except in so far as we have specific reasons to expect a change.'[xxv] Ironically, of course, over the last fifteen years the dominant convention has been to assume that, despite uncertainty-causing innovation, the future will be so like the past that we can trawl data for correlations and patterns that allow us to calculate future risks after all. But this convention was, it turns out, largely baseless. Keynes also emphasised the role of 'animal spirits' in driving our decisions made in uncertainty, and he was under no illusion about the instability as well as dynamism that these spirits can cause. In a fascinating book published last year, George Akerlof and Robert Shiller revisit these insights of Keynes: they show how far markets have been driven by a combination of shifting narratives (including, 'new era stories' that this time is different) and the self-reinforcing emotions of confidence and panic that often fill the void left by uncertainty.[xxvi] Emotions are both vectors and products of particular narratives: they frame and colour our vision and analysis and can affect market behaviour and prices for years at a time.
So what are my conclusions on what market prices tell us? First, that the market may know better than any central authority especially when prices genuinely reflect a plurality of individual perspectives and local contextual knowledge; but in conditions of uncertainty these prices also inevitably reflect the fact that the individuals making transactions rely on socially constructed narratives and learned theories to act as their lanterns to see in the dark. As a result, where markets are dominated by one theory or narrative to the exclusion of others, the expectations reflected in prices may be as much a product of that theory, narrative or convention as of brute reality. Secondly, when the individual expectations reflected in prices are products (in conditions of uncertainty) of individual imagination and intuition rather than mathematical probabilities, they are prone to sudden shifts. And lastly, all market participants (being only human) are prey to emotions that affect their disposition to act. As a result, it is very hard to know ex ante how far prices solve the problem of knowledge or how far instead they reflect private or shared delusions, stories or fleeting emotions.
Let me conclude now with a few words on how we ought to make decisions in conditions of uncertainty instead of relying on the false hopes engendered by the efficient markets hypothesis and quantitative risk analysis. In particular, I would recommend an analytical version of John Keats' 'negative capability'. For Keats, the essence of imaginative genius was the quality of:
'negative capability: that is, when man is capable of being in uncertainties, mysteries, doubts, without any irritable reaching after fact and reason ...[or] ...remaining content with half-knowledge'[xxvii]
I would argue that the entrepreneur or policy-maker trying to reveal hidden aspects of multi-faceted reality or chart her way through the unknown future is likewise better off remaining alert and imaginatively receptive to pointers as they emerge rather than straining to predict the unknowable future or impose her favoured interpretation on the incomplete evidence before her. But such passive imagination is not enough. We also need to experiment with different metaphors and perspectives that can focus our attention on different aspects of reality. It is only by switching our cognitive spectacles regularly that we have much hope of spotting the emergence of novel patterns and trends. Only by experimenting with organic metaphors and non-linear models can we hope, for example, to make sense of network dynamics and self-reinforcing reactions to the crossing of hidden tipping points. Moreover, as David Stark argues in his book, The Sense of Dissonance, it is from the 'generative friction' of using multiple perspectives and evaluative principles that we derive our capacity for innovative thinking. [xxviii] When facing insurmountable uncertainty, disciplined eclecticism and intuitive judgement may be as useful as rational calculation.
And what about government? For Skidelsky, Akerlof and others the prevalence of uncertainty and animal spirits implies a vital role for government. To some extent, this must be true. But it is also worth pointing out that the radical uncertainty implied by innovation and self-reinforcing emotional reactions does more than drive a coach and horses through the assumptions of standard economics and the risk models used by bankers. Uncertainty also imposes limits on the foresight of governments and on the value of government intervention. It may be true that government regulation and intervention can help underpin confidence and reduce uncertainty if done carefully; but if it is seen as arbitrary or manically driven by the latest opinion polls then it can itself cause uncertainty in markets and fear and unease among investors and consumers. In other words, the prevalence of uncertainty and destabilising emotional reactions implies limits on the capacity of governments as well as the likelihood of market failure. We should beware governments bearing five-year plans and insisting on their model of 'best practice' quite as much as economists peddling the efficient markets hypothesis.
© Richard Bronk, Gresham College 2010
[i] Power, Michael, Organised Uncertainty - Designing a World of Risk Management, Oxford University Press, 2007, p. viii.
[ii] Haldane, Andrew G., 'Why Banks Failed the Stress Test', Bank of England, February 2009, p. 4.
[iii] Knight, Frank H., Risk, Uncertainty and Profit, Houghton Mifflin, 1921, p. 233.
[iv] Keynes, John Maynard, The General Theory of Employment, Interest and Money (originally published in 1936), in The Collected Writings of John Maynard Keynes, vol. VII, Macmillan, 1973, pp. 149, 152.
[v] Bronk, Richard, The Romantic Economist - Imagination in Economics, Cambridge University Press, 2009.
[vi] Wordsworth, William, The Prelude (1805), Book II, line 273, ed. E de Selincourt, Oxford University Press, 1960, p. 27.
[vii] Coleridge, Samuel Taylor, Table Talk, 21 September 1830, reprinted in 'The Oxford Authors' Samuel Taylor Coleridge', ed. H.J. Jackson, Oxford University Press, 1985, p. 596.
[viii] Haldane, Andrew G., 'Why Banks Failed the Stress Test', op cit., p. 2.
[ix] Power, Michael, Organised Uncertainty, op cit., p. 74.
[x] Coleridge, Samuel Taylor, 'Lay Sermon addressed to the Higher and Middle Classes' (originally published in 1817), reprinted in Biographia Literaria and Two Lay Sermons, George Bell & Sons, 1898, p. 425f.
[xi] Haldane, Andrew G., 'Rethinking the Financial Network', Bank of England, April 2009.
[xii] Ibid., pp. 24-28.
[xiii] Murdoch, Iris, Sartre - Romantic Rationalist, Vintage, 1999, p. 42.
[xiv] Keynes, John Maynard, The General Theory, op cit., pp. 154-157.
[xv] Soros, George, The Crisis of Global Capitalism - Open Society Endangered, Little Brown, 1998, pp. 47-58.
[xvi] Shackle, G.L.S., Imagination and the Nature of Choice, Edinburgh University Press, 1979, p. 52f.
[xvii] Shackle, G.L.S., Epistemics and Economics - A Critique of Economic Doctrines (originally published in 1972), Transaction Publishers, 1992, pp. xv, 8.
[xviii] Knight, Frank H., Risk, Uncertainty and Profit, op cit., pp. 233, 197-232.
[xix] Keynes, John Maynard, The General Theory, op cit., pp. 161-3, 154.
[xx] Skidelsky, Robert, Keynes - The Return of the Master, Allen Lane, 2009, p. 45.
[xxi] Ibid., p. 88.
[xxii] Knight, Frank H., Risk, Uncertainty and Profit, op cit., p. 311.
[xxiii] Hayek, F.A., 'The Use of Knowledge in Society', The American Economic Review, Vol. 35, No. 4, 1945, pp. 519-530.
[xxiv] Hayek, F.A., 'Competition as a Discovery Procedure', in New Studies in Philosophy, Politics, Economics and the History of Ideas, by F.A. Hayek, University of Chicago Press, 1978, pp. 179-90.
[xxv] Keynes, John Maynard, The General Theory, op cit., pp. 152-4.
[xxvi] Akerlof, George A., and Shiller, Robert, J., Animal Spirits - How Human Psychology Drives the Economy, and Why it Matters for Global Capitalism, Princeton, 2009, pp. 54-6 and passim.
[xxvii] Keats, John, 'Letter to George and Tom Keats', 21 December 1817, extract reprinted in Romanticism - An Anthology, ed. Duncan Wu, Blackwell, 1998, p. 1019.
[xxviii] Stark, David, The Sense of Dissonance - Accounts of Worth in Economic Life, Princeton, 2009, pp. 16-19 and passim.
This event was on Wed, 28 Apr 2010
Gresham College has offered an outstanding education to the public free of charge for over 400 years. Today, Gresham plays an important role in fostering a love of learning and a greater understanding of ourselves and the world around us. Your donation will help to widen our reach and to broaden our audience, allowing more people to benefit from a high-quality education from some of the brightest minds.