Narrative Economics
Economics from a Ricoeurean perspective
'By narrative economics, I mean the study of the spread and dynamics of popular narratives, the stories, particularly those of human interest and emotion, and how these change through time, to understand economics.'
Robert J Shiller, 'Narrative Economics', American Economic Review (2017), 107(4), pp.967-1004
So begins Robert J Shiller's article 'Narrative Economics' which forms the basis of his popular book of the same name. Shiller has a knack for publishing at the opportune moment; Irrational Exuberance (2000), a cautionary tale about the dangers of speculative excess, was released at the height of the great moderation and gave him a reputation as a visionary. Narrative Economics (2019), with its central metaphor of stories as viruses, arrived just months before the first case of SARS-CoV-2 in Wuhan, reinforcing that reputation. It is probably a good idea to keep an eye on Shiller's next project.
Narrative Economics is an interesting book but suffers from the irrepressible urge, not uncommon among economists, to 'quantify first and ask questions later'. It is obvious from the first few pages that Shiller wants to get as quickly as possible to what he calls the epidemiology of narratives (i.e. the dynamics of 'viral' stories), bypassing the difficult conceptual questions that could interrupt the smooth passage from idea to model. Thus when he gets to defining the central concept of his book, this is what Shiller gives us:
'The word narrative is often synonymous with story. But my use of the term reflects a particular modern meaning given in the Oxford English Dictionary: "A story or representation used to give an explanatory or justificatory account of a society, period, etc." Expanding on this definition, I would add that stories are not limited to simple chronologies of human events. A story may also be a song, joke, theory, explanation, or plan that has emotional resonance and that can easily be conveyed in casual conversation. We can think of history as a succession of rare big events in which a story goes viral…"
Robert J Shiller, Narrative Economics, Princeton University Press (2019), p. xii
In other words, narratives are 'memes' in the sense that Richard Dawkins defined them - units of cultural transmission that are replicable and so subject to selective pressures, like viruses. Epidemiology has its own set of tools and naturally Shiller thinks these can be used to add narrative spread/diffusion into economic models of the business cycle and thereby improve forecasting. But is it the case that narratives are simple 'units' that spread like diseases, and which can be modelled using the same mathematics? From one point of view, I suppose the answer is yes. And yet, to hold that point of view to constitute the 'truth' of narrative seems wrong. Is that all that a 'narrative economics' has to offer?
Let's reverse the order of inquiry and ask questions before we quantify, beginning with the most obvious and difficult one - what is a narrative? For a different approach to this question, Shiller could have turned to the work of the French philosopher Paul Ricoeur. I am aware that for some the words 'French' and ‘Philosopher' in the same sentence are enough to conjure visions of monsters. This is because, due to still incompletely understood historical accident, the largely English-speaking world of analytical philosophy separated itself from the largely French-speaking world of continental philosophy sometime in the middle of the last century. The relationship between the two since then resembles a cold war. Ricoeur was unusual as he lived in both worlds quite happily and moving between the two protected him from their characteristic intellectual vices. He wrote on a huge variety of subjects but his most important work is the masterpiece Time and Narrative, a three volumed analysis of the joint problems of narrative epistemology and the phenomenology of time. I consider it a genuine philosophical classic.
As ever the constraints of the blog-form apply; if anything they bite especially hard as Ricoeur's work is both deep and extensive, but a sketch of his main positions is still possible within these limits and may present another way to think about narrative economics. A useful place to start is Ricoeur's statement of his 'working hypothesis' concerning narrative in his essay 'Narrative Time':
'My first working hypothesis is that narrativity and temporality are closely related - as closely as, in Wittgenstein's terms, a language game and a form of life. Indeed, I take temporality to be that structure of existence that reaches language in narrativity and narrativity to be the language structure that has temporality as its ultimate referent. Their relationship is therefore reciprocal.'
Paul Ricoeur, 'Narrative Time', Critical Inquiry , Autumn, 1980, Vol. 7, No. 1, On Narrative (Autumn, 1980), pp. 169-190
This is not to say that narrative is the only way we have to think about time. For example, there is the 'chronological' picture of time as a set of points separated by intervals along the calendar-line. This geometric metaphor allows us to relate events in such a way that one can be said to be 'before' and another 'after'. Chronological time is connected to what Ricoeur calls 'cosmological time' which we measure by reference to the motion of the solar system. Recent discoveries in physics have complicated our understanding of cosmological time by joining space and time together as in the space-time models that underlie the theory of relativity. These examples help explain the point that Ricoeur is making about the unique connection between narrative and temporality. Relativistic physics is a powerful scientific theory that offers deep, often paradoxical insights into the nature of time. Yet for all of its beauty and its minutely validated predictions, it tells us nothing about the experience of being 'inside' time that characterises the human condition. For that we need narrative.
The first part of Time and Narrative Vol I is an analysis of the problem of time from this 'inside' perspective by way of a close reading of St Augustine's Confessions. Ricoeur focuses on the eleventh chapter, which is where Augustine issues his famous sceptical cry - 'What then is time? I know well enough what it is, provided that nobody asks me; but if I am asked what it is and try to explain, I am baffled'. This bewilderment is due to Augustine's inability to bring the two major 'paradoxes of time under rational control. First, the ontological paradox – i.e. in what way can time be said to exist? Second, the measurement paradox – i.e. if we cannot say it exists, how can we measure it? Under the pressure of these twin paradoxes Augustine is compelled to conclude that past, present and future have no substantial reality outside of the mind. Whatever is past no longer exists, whatever is in the future does not yet exist, and the present is just the critical limit where future flows into the past. And so Ricoeur, following Augustine, concludes that our experience of time is characterised by a distentio animi or 'distention of the soul' as it tries to maintain its unity while being stretched between past and future in an 'almost nothing' present.
This 'existential analysis' of time provides the basis for Ricoeur's arguments about the function of narrative. While we have no 'speculative means' to overcome the paradoxes of time they can be resolved at a practical level by the synthetic act of narrative-making. Here Ricoeur follows Aristotle by describing narrative as the form of 'mimesis' concerned with the representation of human action. This then brings in the decisive concept of 'emplotment' which is his creative translation of Aristotle 'muthos'. It is by means of emplotment that the apparent absurdity of time can be given a meaningful shape or ‘configuration’. To fully explain the idea of plot in Time and Narrative would require a long excursion through Ricoeur's theories about the semantics of action and the syntactical features of discourse. The most important thing to emphasise is that through its 'configurational function' the plot enables us to grasp together into a single whole the disparate events of a life, whether fictional or real, so that it can be understood as a unity:
'[Plot] is a mediation between the individual events and a story taken as a whole. In this respect we may say equivalently that it draws a meaningful story from a diversity of events or incidents (Aristotle's pragmata) or that it transforms the events or incidents into a story […] as a consequence an event must be more than just a singular occurrence. It gets its definition from its contribution to the development of the plot. A story too must be more than just an enumeration of events in serial order; it must organise them into an intelligible whole, of a sort that we can always ask what is the 'thought' of this story. In short, emplotment is the operation that draws a configuration out of a simple succession.'
Paul Ricoeur, Time and Narrative, University of Chicago Press (1983), p. 65
The fundamental argument of Time and Narrative across its three volumes is that the activity of narrating is so closely correlated with the temporal character of human experience that it constitutes a 'transcultural form of necessity', by which Ricoeur means that it has something like the status of the ‘pure concepts of the understanding’ that Kant enumerates in the Critique of Pure Reason. These are the a priori concepts through which ‘experience is possible as far as the form of thinking is concerned’. Summarised in a single proposition: 'Time becomes human to the extent that it is articulated through a narrative mode'.
By now you are probably wondering what any of this has to do with Shiller or with economics. There are many ways from Ricoeur's Time and Narrative to economic questions; I will stick to one particular idea. There is a powerful tradition in economics that models the consumer (or producer) as a 'rational maximiser'. In the case of the consumer the thing he is maximising is his utility as represented by a utility function with some specified form. As an example, here is a section from Gerard Debreu's famous monograph The Theory of Value (1959), a landmark in mathematical economics:
Given a price-wealth pair (p, w) in Si, the ith consumer chooses, in the non-empty set y,(p, w), a consumption z, which is optimal according to his preferences, i.e., a greatest element for the preference pre-ordering. If there is a utility function ui one can also say that he chooses a maximiser of u, on y,(p, w) (in this case "preference satisfaction" is therefore synonymous with "utility maximization"). Doing this amounts to selecting the quantities of each good or service he will consume, and the quantities of each type of labour he will produce (at each date and location) which form a possible consumption plan optimal for his limited wealth.
G Debreu, Theory of Value: An Axiomatic Analysis of Economic Equilibrium, Yale University Press (1959), p. 65
If we start with some Ricoeurian ideas about the centrality of narrative to human experience, we can keep the idea of the human agent as a maximiser but will probably have to change the function that he is maximising. In a later book, Oneself as Another (1990), Ricoeur proposes that the primary ethical motive underlying our choices is to create and preserve the 'identity of character' or 'the unity of a life considered as a temporal totality which is itself singular and distinguished from all others'. Using Ricoeur’s language, we want to fit our 'life plans' - 'those vast practical units that make up professional life, family life, leisure time' - into the 'global project of our existence' in a way that is coherent (or ‘concordant’). We are thus less interested in maximising utility per se than we are in maximising the degree of coherence that our life presents to us when understood as a ‘global’ narrative.
This is a fundamentally different way of looking at economic decisions and may go some way to explaining why people make choices that often seem irrational from the perspective of their strict economic pay-offs. The 'successive' or 'episodic' nature of narrative has a further implication for economics. Because narrative unfolds over time a Ricoeurean agent cannot be modelled following the standard equilibrium analysis that require the consumer to 'choose a consumption plan made now for the whole future' (Debreu). To borrow again from Ricoeur, his analysis forces us to ask whether the choices made by human agents can be fully translated, without loss, into the 'a-chronology of models'.