by Kate Raworth
With such ill-fitting models dominating macroeconomic analysis, some big-name insiders began to critique the very theories that they had helped to legitimise. Robert Solow, known as the father of neoclassical economic growth theory and long-time collaborator of Paul Samuelson, became an outspoken critic, first in his 2003 speech bluntly entitled ‘Dumb and Dumber in Macroeconomics’, then in analyses that mocked the theory’s stringent assumptions.6 The general equilibrium model, he pointed out, in fact depends upon there being just one single, immortal consumer-worker-owner maximising their utility into an infinite future, with perfect foresight and rational expectations, all the while served by perfectly competitive firms. How on earth did such absurd models come to be so dominant? In 2008, Solow gave his view:
I am left with a puzzle, or even a challenge. What accounts for the ability of ‘modern macro’ to win hearts and minds among bright and enterprising academic economists? … There has always been a purist streak in economics that wants everything to follow neatly from greed, rationality, and equilibrium, with no ifs, ands, or buts … The theory is neat, learnable, not terribly difficult, but just technical enough to feel like ‘science.’ Moreover it is practically guaranteed to give laissez-faire-type advice, which happens to fit nicely with the general turn to the political right that began in the 1970s and may or may not be coming to an end.7
One thing that is clearly coming to an end is the credibility of general equilibrium economics. Its metaphors and models were devised to mimic Newtonian mechanics, but the pendulum of prices, the market mechanism, and the reliable return to rest are simply not suited to understanding the economy’s behaviour. Why not? It’s just the wrong kind of science.
No one made this point more powerfully than Warren Weaver, the director of natural sciences at the Rockefeller Foundation, in his 1948 article, ‘Science and Complexity’. Looking back over the last three hundred years of scientific progress, while simultaneously looking forward at the challenges facing the world, Weaver clustered together three kinds of problems that science can help us to understand. At one extreme lie problems of simplicity, involving just one or two variables in linear causality – a rolling billiard ball, a falling apple, an orbiting planet – and Newton’s laws of classical mechanics do a great job of explaining these. At the other extreme, he wrote, are problems of disordered complexity involving the random movement of billions of variables – such as the motion of molecules in a gas – and these are best analysed using statistics and probability theory.
In between these two branches of science, however, lies a vast and fascinating realm: problems of organised complexity, which involve a sizeable number of variables that are ‘interrelated in an organic whole’ to create a complex but organised system. Weaver’s examples came close to asking the very questions that Newton’s apple failed to prompt. ‘What makes an evening primrose open when it does? Why does salt water fail to satisfy thirst? … Is a virus a living organism?’ He noted that economic questions came into this realm, too. ‘On what does the price of wheat depend? … To what extent is it safe to depend on the free interplay of such economic forces as supply and demand? … To what extent must systems of economic control be employed to prevent the wide swings from prosperity to depression?’ Indeed, Weaver recognised that most of humanity’s biological, ecological, economic, social and political challenges were questions of organised complexity, the realm that was least understood. ‘These new problems, and the future of the world depends on many of them, require science to make a third great advance,’ he concluded.8
That third great advance got under way in the 1970s when complexity science – which studies how relationships between the many parts of a system shape the behaviour of the whole – began to take off. It has since transformed many fields of research, from the study of ecosystems and computer networks to weather patterns and the spread of disease. And although it is all about complexity, its core concepts are actually quite simple to grasp – meaning that, despite our instincts, we can all learn, through training and experience, to be better ‘systems thinkers’.
A growing number of economists are thinking in systems too, making complexity economics, network theory, and evolutionary economics among the most dynamic fields of economic research. But, thanks to the lasting influence of Jevons and Walras, most economics teaching and textbooks still introduce the essence of the economic world as linear, mechanical and predictable, summed up by the market’s equilibrating mechanism. It’s a mindset that will leave future economists deeply ill-equipped to handle the complexity of the contemporary world.
In a playful ‘look back from 2050’ the economist David Colander recounts that, by 2020, the majority of scientists – from physicists to biologists – had already realised that complexity thinking was essential for understanding much of the world. Economists, however, were a little slower on the uptake and it was not until 2030 that ‘most economic researchers believed that the economy was a complex system that belonged within complexity science’.9 If his history of the future should turn out to be right, it may well be too late. Why wait until 2030 when we can ditch the ill-chosen metaphors of Newtonian physics and get savvy with systems now?
The dance of complexity
At the heart of systems thinking lie three deceptively simple concepts: stocks and flows, feedback loops, and delay. They sound straightforward enough but the mind-boggling business begins when they start to interact. Out of their interplay emerge many of the surprising, extraordinary and unpredictable events in the world. If you have ever been mesmerised by the sight of thousands of starlings flocking at sunset – in a spectacle poetically known as a murmuration – then you’ll know just how extraordinary such ‘emergent properties’ can be. Each bird twists and turns in flight, using phenomenal agility to stay a mere wingspan apart from its neighbours, while tilting as they tilt. But as tens of thousands of birds gather together, all following these same simple rules, the flock as a whole becomes an astonishing swooping, pulsing mass against the evening sky.
So what is a system? Simply a set of things that are interconnected in ways that produce distinct patterns of behaviour – be they cells in an organism, protestors in a crowd, birds in a flock, members of a family, or banks in a financial network. And it is the relationships between the individual parts – shaped by their stocks and flows, feedbacks, and delay – that give rise to their emergent behaviour.
Stocks and flows are the basic elements of any system: things that can get built up or run down – just like water in a bath, fish in the sea, people on the planet, trust in a community, or money in the bank. A stock’s levels change over time due to the balance between its inflows and outflows. A bathtub fills or empties depending on how fast water pours in from the tap versus how fast it drains out of the plughole. A flock of chickens grows or shrinks depending on the rate of chicks born versus chickens dying. A piggy bank fills up if more coins are added than are taken away.
If stocks and flows are a system’s core elements, then feedback loops are their interconnections, and in every system there are two kinds: reinforcing (or ‘positive’) feedback loops and balancing (or ‘negative’) ones. With reinforcing feedback loops, the more you have, the more you get. They amplify what is happening, creating vicious or virtuous circles that will, if unchecked, lead either to explosive growth or to collapse. Chickens lay eggs, which hatch into chickens, and so the poultry population grows and grows. Likewise, in the vengeful tit-for-tat of playground fights, a single rough shove can soon escalate into a full-blown bust-up. Interest earned on savings adds to those savings, increasing future interest payments, and so wealth accumulates. But reinforcing feedback can lead to collapse too: the less you have, the less you get. If people lose confidence in their bank and withdraw their savings, for example, it will start to run out of cash, deepening the loss of confidence and leading to a run on the bank.
If reinforcing feedbacks are what make a system move, then balancing feedbacks are what stop it from
exploding or imploding. They counter and offset what is happening, and so tend to regulate systems. Our bodies use balancing feedbacks to maintain a healthy temperature: get too hot and your skin will start sweating in order to cool you down; get too cold and your body will start shivering in an attempt to warm itself up. A household’s thermostat works in a similar way to stabilise room temperature. And in a playground scuffle, someone is likely to step in and try to break it up. In effect, balancing feedbacks bring stability to a system.
Complexity emerges from the way that reinforcing and balancing feedback loops interact with one another: out of their dance emerges the system’s behaviour as a whole, and it can often be unpredictable. The simplest depiction of the ideas at the heart of systems thinking is a pair of feedback loops, and the one shown here tells a simple story of chickens, eggs, and crossing the road.10
Each arrow shows the direction of causation and comes with a plus or minus sign. A plus sign indicates that the effect is positively related to the cause (more chickens result in more attempted road crossings, for example) while a minus sign stands for the reverse (more attempted road crossings result in fewer chickens). Each pair of arrows creates a loop, labelled R if it is reinforcing and B if it is balancing. On the left, more chickens lay more eggs that hatch into more chickens: a reinforcing loop. On the right, more chickens make more attempted road crossings, which results in fewer chickens: a balancing loop. When both feedback loops are in play in a highly simplified system like this one (assuming there is at least one rooster in the flock and no shortage of grain), what might happen to the size of the poultry population over time? Depending on the relative strength of the two loops – the rate at which chickens produce chicks versus the rate at which chickens get hit – the flock might grow exponentially, collapse, or even come to oscillate continuously around a stable size if there is a significant delay between chicks hatching and their attempting to cross the road.
Feedback loops: the fundamentals of complex systems. Reinforcing feedback (R) amplifies what is happening, while balancing feedback (B) counters it. Their interaction creates complexity.
Delays such as this – between inflows and outflows – are common in systems and can have big effects. Sometimes they bring useful stability to a system, allowing stocks to build up and act as buffers or shock absorbers: think energy stored in a battery, food in the cupboard, or savings in the bank. But stock–flow delays can produce system stubbornness too: no matter how much effort gets put in, it takes time to, say, reforest a hillside, build trust in a community, or improve a school’s exam grades. And delay can generate big oscillations when systems are slow to respond – as anyone knows who has been scalded then frozen then scalded again while trying to master the taps on an unfamiliar shower.
It is out of these interactions of stocks, flows, feedbacks and delays that complex adaptive systems arise: complex due to their unpredictable emergent behaviour, and adaptive because they keep evolving over time. Beyond the realm of starlings and chickens, bathtubs and showers, it soon becomes clear just how powerful systems thinking can be for understanding our ever-evolving world, from the rise of corporate empires to the collapse of ecosystems. Many events that first appear to be sudden and external – what mainstream economists often describe as ‘exogenous shocks’ – are far better understood as arising from endogenous change. In the words of the political economist Orit Gal, ‘complexity theory teaches us that major events are the manifestation of maturing and converging underlying trends: they reflect change that has already occurred within the system’.11
From this perspective, the 1989 fall of the Berlin Wall, the 2008 collapse of Lehman Brothers and the imminent collapse of the Greenland ice sheet have much in common. All three are reported in the news as sudden events but are actually visible tipping points that result from slowly accumulated pressure in the system – be it the gradual build-up of political protest in Eastern Europe, the build-up of sub-prime mortgages in a bank’s asset portfolio, or the build-up of greenhouse gases in the atmosphere. As Donella Meadows, one of the early champions of systems thinking, put it, ‘Let’s face it, the universe is messy. It is nonlinear, turbulent, and chaotic. It is dynamic. It spends its time in transient behaviour on its way to somewhere else, not in mathematically neat equilibria. It self-organises and evolves. It creates diversity, not uniformity. That’s what makes the world interesting, that’s what makes it beautiful, and that’s what makes it work.’12
Complexity in economics
The realisation that economics needs to embrace dynamic analysis is by no means a recent one. Over the past 150 years, economists of all stripes tried to break away from imitating Newtonian physics, but their efforts were all too often steam-rolled by the dominance of equilibrium theory and its satisfyingly neat equations. Jevons himself had a hunch that economic analysis should be dynamic but, lacking the mathematics to do it, he settled for comparative statics, which compares snapshots of two points in time: it was an unfortunate compromise because it led him away from, rather than towards, the insight he ultimately sought.13 In the 1860s, Karl Marx described how the relative income shares of workers and capitalists would continually rise and fall, due to self-perpetuating cycles of output and employment.14 By the end of the nineteenth century, Thorstein Veblen was criticising economics for being ‘helplessly behind the times in not being evolutionary’ and therefore unable to explain change or development,15 while Alfred Marshall argued against mechanical metaphors and, instead, for seeing economics as ‘a branch of biology, broadly interpreted’.16
Twentieth-century attempts to recognise the economy’s inherent dynamism were likewise made by deeply opposing schools of thought but even they couldn’t dislodge equilibrium thinking. In the 1920s John Maynard Keynes critiqued the use of comparative statics, pointing out that it is precisely what happens in between those snapshots of economic events that is of greatest interest. ‘Economists set themselves too easy, too useless a task,’ he wrote, ‘if in tempestuous seasons they can only tell us that when the storm is long past the ocean is flat again.’17 In the 1940s, Joseph Schumpeter drew on Marx’s insights into dynamism to describe how capitalism’s inherent process of ‘creative destruction’, through continual waves of innovation and decline, gave rise to business cycles.18 In the 1950s, Bill Phillips created his MONIAC precisely with the aim of replacing comparative statics with system dynamics, complete with the time lags and fluctuations that can be observed as water flows into and out of tanks. In the 1960s Joan Robinson lambasted equilibrium economic thinking, insisting that, ‘a model applicable to actual history has to be capable of getting out of equilibrium; indeed it must normally not be in it’.19 And in the 1970s, the father of neoliberalism, Friedrich Hayek, decried the economist’s ‘propensity to imitate as closely as possible the procedures of the brilliantly successful physical sciences – an attempt which in our field may lead to outright error’.20
So let’s finally heed their collective advice, push equilibrium thinking to one side, and start to think in systems instead. Imagine pulling the iconic supply and demand curves out of their rigid criss-cross and twisting them into a pair of feedback loops. At the same time, drop the economist’s beloved notion of ‘externalities’, those incidental effects felt by people who were not involved in the transactions that produced them – like toxic effluent that affects communities living downstream of a river-polluting factory, or the exhaust fumes inhaled by cyclists biking through city traffic. Such negative externalities, remarks the ecological economist Herman Daly, are those things that ‘we classify as “external” costs for no better reason than because we have made no provision for them in our economic theories’.21 The systems dynamics expert John Sterman concurs. ‘There are no side effects – just effects,’ he says, pointing out that the very notion of side effects is just ‘a sign that the boundaries of our mental models are too narrow, our time horizons too short’.22 Due to the scale and interconnectedness of the global economy, many economic ef
fects that were treated as ‘externalities’ in twentieth-century theory have turned into defining social and ecological crises in the twenty-first century. Far from remaining a peripheral concern ‘outside’ of economic activity, addressing these effects is of critical concern for creating an economy that enables us all to thrive.
From this vantage point – counter-intuitive though it may sound – equilibrium economics actually turns out to be a form of systems analysis, just an extremely limited one. It get the results it seeks by imposing severely restrictive assumptions about how market systems behave – assumptions including perfect competition, diminishing returns, full information, and rational actors – so that no errant effects get in the way of the price mechanism’s ability to act as the balancing feedback loop that restores market equilibrium. Think of it in terms of starlings: what restrictions would you have to impose on a large flock of these birds if you wanted to make sure that they all stayed still? You could place each bird in its own narrow little coop and shut them all away in a dark, quiet room: that should encourage them to stay put. But don’t expect the flock to behave like that once you remove these unnatural confines and release them into the air. They will twist and turn, putting on an extraordinary aerial display of a complex system in action. So it is with economic actors trapped in the narrow confines of an equilibrium model: when all the restrictive assumptions are in place, they will indeed behave as required. But remove those assumptions – enter the real world – and all havoc could break loose. It often does, of course, in the boom-to-bust of financial crash, in the rise of the 1%, and in the tipping points of climate change.
