By Paul Ormerod
Modern economic theory was first set out on a formal basis in the late 19th century. While there have been many developments since then, at heart the view in economics of how the world operates remains the same. Mainstream economic theory is essentially concerned with how decisions are made by individuals, what information is gathered and how it is used by decision-makers. The intellectual basis of most social and economic policy in the western world is provided by this conventional economic theory. A whole range of activities in state bureaucracies, such as forecasts, policy evaluation, cost–benefit analysis and the design of regulation, stem directly from mainstream economics.
All scientific theories, even quantum physics, are approximations to reality. Developing theories involves making assumptions and simplifications to enable us to better understand problems. A key feature of a good theory, therefore, is that its assumptions are a reasonable description of the real world.
In the early 21st century, just as in the late 19th, economics in general makes the assumption that individuals operate autonomously, isolated from the direct influences of others. A person has a fixed set of tastes and preferences; when choosing from a set of alternatives, he or she compares the attributes of those alternatives and selects the one which most closely corresponds to his or her preferences. At first sight, this may seem quite reasonable, or even ‘rational’, as economists describe this theory of behaviour. If I am interested in buying a product which many people want, I may have to pay a high price. So the choices other people make affect me indirectly through the workings of the market. My preferences, however, remain unaltered, according to this conventional view of economics.
Get Evonomics in your inbox
There is a serious problem with this assumption that individuals operate in isolation from each other, that their preferences are not affected directly by the decisions of others. The social and economic worlds of the 21st century are simply not at all like this. In the real world we are far more aware than ever before of the choices, decisions, behaviours and opinions of other people. In 1900, not much more than 10 percent of the world’s population lived in cities. Now, for the first time in human history, more than half of us do, in close, everyday proximity to large numbers of other people. In addition, since the end of the 20th century the internet has revolutionised communications in a way not experienced since the invention of the printing press 500 years earlier.
So the assumption that people make choices in isolation, that they do not adopt different tastes or opinions simply because other people have them, is no longer sustainable. Perhaps – and it is a big ‘perhaps’ – a century or so ago this was a reasonable assumption to make, but no longer.
The choices people make, their attitudes and their opinions are influenced directly by others and the medium across which this influence spreads is social networks. Commonly, social networks are thought of as purely a web-based phenomenon: sites such as Facebook, Twitter and MySpace. These online social networks indeed can influence behaviour, but it is real-life social networks – such as family, friends, colleagues – that are even more important in helping us shape our preferences and beliefs, what we like and what we do not like.
The fact that a person can and often does decide to change his or her preferences simply on the basis of what others do is known in economics as network effects. Also called network externalities or demand-side economies of scale, network effects pervade the modern world.
Network effects have in fact been pervasive throughout human history. A crucial feature of human behaviour is our propensity to copy or imitate the behaviours, choices and opinions of others. We can see it, for example, in the fashions in pottery in the Middle Eastern Hittite Empire of three and a half millennia ago. And we can see it today in the behaviour of traders on financial markets, where the propensity to follow the herd can lead all too easily to the booms and crashes experienced by economies around the world.
Networks are especially important in finance. When, in September 2008, Lehman Brothers went bankrupt, it precipitated a crisis that almost led to a total collapse of the world economy and a repeat of the Great Depression of the 1930s. It was precisely because Lehman was connected into a network of other banks that the situation was so serious. Lehman’s bankruptcy could easily have led to a cascade of bankruptcies across the global financial network, initially in those institutions to which Lehman owed money, and then spreading wider and wider across the entire network as more and more institutions became exposed. Incredibly, neither the systems of financial regulations which were in place, nor the thinking of mainstream economics that influenced policy so strongly, took any account of the possibility of such a network effect. Ironically, policymakers and the financial establishment thought that risk could be mitigated by spreading it across the system in the form of securitisation and the slicing up of risky assets. They misunderstood completely the dynamics of financial networks and the possibility that such networks would not reduce risk but instead would trigger uncertainty and upheaval.
A world in which network effects are a driving force of behaviour is completely different from the world of conventional economics, in which isolated individuals carefully weigh up the costs and benefits of any particular course of action. A world in which network effects are important is a much more realistic description of the human social and economic worlds which actually exist in the 21st century. Incentives, of course, have not disappeared as a driver of human behaviour: it is still the case that if, say, Pepsi raises its price compared to Coca-Cola, more Coke and less Pepsi will be sold. This is the world that conventional economic theory describes. It is not wrong, but it is often misleading. It offers only a very partial account of how decisions are made in reality, where network effects can have far greater influence on behaviour than incentives. Network effects, in fact, can completely swamp the impact of incentives, leading to very different outcomes to the ones intended by those who altered the incentives, whether they are companies or public policymakers.
Network effects require policymakers, whether in the public or corporate spheres, to have a markedly different view of how the world operates. They make successful policy much harder to implement and they help explain many of the failures of policies that are based on the view that incentives, rather than network effects, are the key drivers of behaviour. Understanding the influence of network effects and harnessing our knowledge of how they work in practice, however, opens up the possibility of far more effective and successful economic policies.
Why we need a new intellectual foundation of policy
We might reasonably reflect that we are very much better off than we were in, say, the middle of the 20th century. Over this period, we have had a great deal of state activity, of public policy interventions in both social and economic problems based on the model of economically rational agents: that agents respond solely to incentives. Network effects and copying are entirely absent from this model. So why do we need a new perspective on policy at all? Surely we have done well using the old model, especially when boosted by the addition of the late 20th century insights into asymmetric information and the principle of ‘market failure’?
A distinguishing feature of the social and economic history of the second half of the 20th century is the enormous rise in the role of the state throughout the western world. Gradually, many of the functions previously within the domain of the third or private sectors have been embraced within the public sector. President Roosevelt’s New Deal in the US in the 1930s was bitterly denounced by critics at the time as being nothing less than socialism. But the percentage share of the whole economy accounted for by the spending of the Federal government then was not much more than half of what it was under Ronald Reagan 50 years later. Likewise, the most avowedly socialist government in the history of the UK was that of Clement Attlee from 1945 to 1951. Yet the share of the public sector in the economy as a whole under Attlee was less than it was during the government of Margaret Thatcher (1979–1990), renowned for her robust approach to the privatisation of state activities.
The intellectual underpinning of the burgeoning activity of the state has been provided by mainstream economics. Paradoxically, a theoretical construct which purports to establish the efficiency of the free market has justified an enormously enhanced role for the state. It is not just the sheer size of the public sector, but the range of private activities which the state now tries to influence or control: either through direct regulation or through exhortations to avoid behaviours deemed inappropriate by civil servants, such as those which lead to obesity or the consumption of anything more than small amounts of alcohol. Generations of policymakers have been raised to have a mechanistic view of the world, and a checklist mentality: to achieve a particular set of aims, draw up a list of policies, and simply tick them off. It is a comforting environment in which to live, being seemingly dependable, predictable and controllable.
The concept of ‘market failure’, at first sight a critique of free market economics, has provided powerful backing to state intervention. If markets, for whatever reason, are unable to function in practice as the theory suggests they should, then regulation, taxes, incentives of all shapes and forms, are justified. They are justified in order to make the imperfect world conform to the perfect one of economic theory. Economists slip all too easily into the attitude that their core theory does not merely purport to describe how the world actually is, it is a prescription for how the world ought to be.
The world view of free market economic theory is precisely one in which rational agents are able to make optimal decisions and achieve the best possible outcome in any particular set of circumstances. And so behaviour can be influenced by the appropriate set of incentives selected by the authorities. Indeed, we see a vast array of taxes, subsidies and benefits, all aimed at achieving precise and detailed outcomes. And where there are obstacles to agents making the best choice, where there is ‘market failure’, the clever, rational planner can intervene to ensure that the world works as the theory deems it should do.
We have now had over 60 years of this vision. It is fundamentally different from anything seen before in the western world, except during the two world wars. And yet, the stark fact is that the combination of large-scale state activity and a mechanistic intellectual approach to policymaking has simply not delivered anything like the success that the founding fathers of the post-second world war social settlement imagined would be the case. Deep social and economic problems remain. For example, both the average rate of unemployment and the range within which it varies are scarcely any different in the six decades since the end of the second world war to the same period preceding it. If policy planners were supposed to achieve anything, then surely – in the wake of the massive unemployment that was the scourge of the west in the early 1930s – it was very low levels of unemployment. To be fair, the maximum rates of unemployment in the west have not hit the heights of the Great Depression, but unemployment has consistently remained a serious problem and in 2012 stands at an alarmingly high level in many countries.
Taking a long-term view, averaging over decades, the unemployment rates are very similar in the pre- and post-second world war periods. In the US, the pre-war average was 7 per cent compared to just under 6 per cent post-war; in the UK the two averages are virtually identical at around 5.5 per cent; while in Germany, for example, the average unemployment rate since the second world war, at just over 5 per cent, is even higher than the pre-war average of some 4 per cent.
Other social and economic problems also remain deeply rooted. Comparing crime rates over time is difficult, but despite sharp falls since the mid-1990s in both the US and Britain, for example, crime is everywhere much higher now than it was in 1950. Meanwhile, the distribution of income and wealth has widened dramatically, while rational planning and clever regulation designed to cope with ‘market failure’ did not prevent the biggest economic recession since the 1930s from taking place in 2008/09.
However, the principal cause of the failure of what we might describe as the post-war western model to achieve its objectives is not the size of the state but the intellectual framework in which it operates. The differences between the centre-right and centre-left within this model have been of merely second-order importance. Both the main parts of the political spectrum have embraced not only a much greater role for the state than obtained before the second world war, but have shared this same intellectual vision. At heart, from this perspective the world is seen as a machine, admittedly a complicated one, but one that can be controlled with the right pressure on this button, just the right amount of pull on that lever. It is a world in which everything can be quantified and targets can not only be set, they can be achieved thanks to the cleverness of experts. But the world is simply not like this. It is a much more complex, much less controllable place than ‘rational’ planners believe. Policy is very difficult to get right.
The belief that clever people, with sufficient thought, really can be social engineers and design the perfect society is very deeply embedded. Just as real engineers can design bridges that work exactly as intended, so the vision of society and the economy as machines encourages policymakers to take the same view of their ability to design human behaviour. But it is no longer relevant, if it ever was, to most aspects of human social and economic behaviour. This is why the network effects view of behaviour is so challenging. I have heard frequent arguments along the lines: this is all very well, these networks seem very clever, but you lack clear guidelines about what we should actually do to solve a problem. If we use the economically rational approach, we know what to do.
The latter point is an obvious non sequitur. The economic rational agent model is indeed capable of providing policymakers with an exact answer to a problem: in order to achieve X, do Y. But all too often, doing Y leads to Z, or even to what we might call ‘minus X’ – in other words, the complete opposite of what was intended. During 2011, for example, there was constant concern about the state of Europe’s economies, and the future of both the euro and the eurozone. Periodically, the French president or the German chancellor or the head of the European Commission made a statement intended to calm the markets, or the European Central Bank intervened in the bond market with the same intention in mind. But instead of recovering, the markets often fell further. This way of thinking about policy does not provide control, merely the illusion of control.
These criticisms do not apply simply to the centre-left, where the often uncritical elevation of public bureaucracy into a ‘good thing’ has become a hallmark. A sharp distinction also needs to be made between the network approach to understanding both society and the economy and, on the other hand, the Chicago, free market approach beloved of the centre-right. Ironically, this latter modus operandi is the mirror image of the ‘clever planner’ concept of policymaking. So, if the correct set of prices – read incentives – can be put in place, all markets will operate efficiently, and no resources will be left unused; all markets will clear. The role of the state is minimal. But how are these prices ever to be discovered? Economic theory gives no guidance here at all, and is forced to rely on a mythical creature called the Auctioneer to perform this task. It is merely a short step from this vision of the world to the bureaucrat and his or her belief that the right benefit, the right tax rate, the right regulation can be set in order to achieve any desired aim.
In any event, the network view of the world inherently gives rise to the concept of collective action. If a set of values spreads across a network, the behaviour of the individual component parts is altered by these emergent, collective values. The agents in the network are not isolated individuals, but operate in society and have their behaviour, at least in part, shaped by society.
Take, for example, the issue of drink-driving. In most social circles today, driving after drinking substantial amounts of alcohol is the subject of strong disapproval. But this was not always the case. When Barbara Castle introduced drink-driving legislation in the late 1960s, it met with strong resistance and was widely ignored. Many of the late-night heavy drinkers in my uncle’s pub on the moors above Bolton were the police themselves. Gradually, however, a different set of attitudes spread across social networks and drink-driving has become very much less widespread than it was then. Even though the penalties are severe, the chances of being caught remain very low. It is the social norms that have emerged rather than the legislation which keep drink-driving incidents down. A contrast is provided by attitudes to the speed limit, especially on motorways and fast dual carriageways. The 70mph limit has very little social acceptance and is routinely ignored, again despite penalties if caught. On the M40 motorway, for example, more vehicles travel at speeds in excess of the legal speed limit than at 70mph or below.
So, sometimes traditional legislation works – and its impact is boosted by emergent social norms – and sometimes it does not. But the key feature of both the drink-driving and the speed limit examples is that it is the social norm, emerging across networks, which is critical. This collective feature of the network exercises a powerful influence on the behaviour of the individual agents within the network.
The recognition of the fundamental importance of networks for outcomes in the modern social and economic worlds does not mean that governments are powerless. Instead it calls for smarter government rather than no government. It almost certainly means fewer state bureaucrats, working in an outdated intellectual framework, searching for the elusive silver bullet which is guaranteed to solve a problem.
The silver bullet of this approach is that there are no silver bullets. Instead, we need to rely much more on the processes of experimentation and discovery. A key influence on behaviour in many social and economic contexts is the prevailing social norm in the relevant network, which emerges from the interactions of the individuals who comprise the network. But there are no levers, no magic buttons to press, which will guarantee that social norms can be altered in ways which the policymaker desires. We can only discover what works by experiment.
This does not mean that we are operating in the dark, that the success or otherwise of a policy is merely a matter of chance. The more knowledge we have of how people are connected on the relevant network, of who might influence whom and when, the more chance a policy has of succeeding. Much of this knowledge is held at decentralised levels in tacit form, a form which is hard or even impossible to codify. But it is crucial to how most social and economic systems work in practice.
Our current political institutions are to a large extent based on the vision of society and the economy operating like machines, populated by economically rational agents. This view of the world leads to centralised bureaucracies and centralised decision-making. We live in a society where decisions are made through several layers of bureaucracy, in both the public and private sectors. On the whole, this leads to decisions that are insensitive to local (micro) conditions, and which are insensitive to society as it changes.
A lack of both resilience and robustness is a characteristic feature of such approaches to social and economic management. Structures, rules, regulations, incentives are put in place in the belief that a desired outcome can be achieved, that a potential crisis can be predicted and forestalled by such policies. As the recent financial crisis illustrates only too well, this view of the world is ill-suited to creating systems which are resilient when unexpected shocks occur, and which exhibit robustness in their ability to recover from the shock. The focus of policy needs to shift away from prediction and control. We can never predict the unpredictable. Instead, we need systems which exhibit resilience and robustness together with the ability to adapt and respond well to unpredictable future events.
Adapted from Complex New World: Translating new economic thinking into public policy, published by the Institute for Public Policy Research (IPPR).
2016 July 17