Ian Goldin is a former vice-president of the World Bank and advisor to President Nelson Mandela. He is now director of the Oxford Martin School and professor of globalisation and development at the University of Oxford. In his new book, "The Butterfly Defect", co-written with Mike Mariathasan, Goldin explores both the promise and the perils of globalisation. "We are more tightly linked than ever before," they write.
That unprecedented integration of economic and other systems has brought with it previously unimaginable benefits to substantial portions of humanity. But it also exposes us to "systemic risks" of a kind we've not faced before, and which politicians, at both the national and global level, need to try to understand—and quick. They could do worse than read "The Butterfly Defect," though its analysis of the risks that metastasise in fiendishly complex global systems will probably keep them awake at night.
I met Goldin in London earlier this week and began by asking him what he thinks the principal gains and benefits of globalisation have been.
IG: I think the last 25 years or so—the period since the fall of the Berlin Wall—have seen the most remarkable progress that the world has ever known in many dimensions of human wellbeing. Life expectancy, the reduction of poverty, quality of life, nutrition etc. If you look at this period in the long sweep of history, it will come to be seen as a period in which humanity leapt ahead in many dimensions. I attribute that largely to the coming down of walls and integration, which is how I define globalisation. It’s the flows between societies, principally the flow of ideas, that has lead to these changes.
Now, with that also comes great risk. For example, climate change, pollution, the destruction of biodiversity and so on are negative spillovers [of globalisation]. Then there is a series of new, emergent risks. I see the financial crisis of 2008 as one of the clearest exemplars of these new kinds of risk. It’s what I call the “butterfly defect"—small perturbations somewhere leading to these cascading, amplified shocks which can be deeply destabilising. Because there’s a mismatch between the integration processes and global governance, the risks are growing. You can’t manage globalisation at the local or national level alone. It needs management at those levels, and there’s a lot you can do, but you’re not going to solve climate change, pandemics, cyberterrorism or the problems of finance in one country. And that’s the tension.
JD: Central to your analysis of globalisation is the notion of “complexity”. Your claim is that complex systems incubate “systemic risks”.
Right. There’s a vast body literature that comes from physics, in the first instance, about how systems evolve and about how they can be made more robust and resilient. It’s also about how the nodes and networks in systems need to be understood in order to understand system stability. It is also important to understand attribution—cause and effect—in a system. And unless you model complex systems, it’s very difficult to know where the beginning and the end is. They become like entangled webs.
Now, part of the problem that politicians face is this attribution problem. They can no longer say [to voters], “I know the cause of your problems.” Now the problems they face come from beyond [national] borders. Bookshelves have been filled already with books about the financial crisis. But no one is going to be able to prove that it was caused by a, b or c. It’s too complex. Let’s take an example: if you think that systemic risk is likely to be propagated through key nodes—one bank, like Lehman Brothers, or one airport, like Heathrow, or one port, like Houston—you can design competition policy to ensure that no single place is too big to fail. If you think that systemic risk in transport hubs is a crucial factor in thinking about the stability of the airline network in the UK, for example, you might take a different view about the growth of Heathrow.
This isn’t the first age of globalisation, of course. But you do think technological progress and innovation makes the present age of globalisation distinctive, and that it makes the attendant risks distinctively threatening don't you?
Yes. This is a very different kind of globalisation to the previous waves that have gone before. This one is different for a number of reasons. One is the number of countries involved. Previous waves of globalisation tended to involve small sets of countries—for example, Britain globalising with its colonies. This wave is much broader because it’s across multiple dimensions. Virtually everything you can think of is being globalised. Then there’s the number of countries involved. There’s really only one country left in the world that’s not integrated [into the global system], and that’s North Korea. So whereas 20 or 30 years ago, half the world’s population was not connected and open, today maybe one per cent is.
The third important factor is the speed of integration, which is related to technology. It’s not only physical connectivity, it’s virtual connectivity. So that’s fundamentally different. Speed of transmission has changed.
You write that one of the features of the contemporary landscape is the unravelling of the “classical distinction between risk and uncertainty”. What do you mean by that?
Risk is something that you can quantify. It has a probability associated with it. Uncertainty is different: you cannot assign a probability to it. What I call “uncertainty” is not what Nicholas Taleb means when he talks about “black swan” events. Those are exogenous events. What I’m talking about are long-tail events. They’re different because they’re endogenous, they’re caused by us, by globalisation, and we could stop them happening if we understood complex systems well enough and governed them properly. One of the problems in unstable systems is not only how problems amplify and cascade from a single source, but also how they can hop between sectoral boundaries—how a pandemic, say, can lead to a financial crisis, or how a drought can lead to a war. I define “systemic risks” as risks that spill over borders. Borders between countries, but also borders between sectors.
Let’s turn now to the financial sector, specifically. I was struck by what you were just saying about endogenous risks and uncertainties. Is there an affinity between your argument here and the work of Hyman Minsky on financial instability?
I think there is. And also with the work of Keynes and Robert Shiller on herding behaviour, “animal spirits” and so on. These [financial] systems are inherently unstable and the market is unstable. They also tend to monopolistic power, because companies can pay for the lobbying which stops new entrants to the market. That’s why you need competition policy.
Among the many failures that led to the blow-up of the global financial system in 2008 was, you suggest in the book, a failure of the economics profession. You mentioned regulatory failures and failures of competition law just now. There was a fairly sustained assault on competition law, antitrust legislation and so on waged for many years by economists at the University of Chicago—one thinks of the work of Ronald Coase, Robert Bork and other members of the “Law and Economics” school.
Right. It’s not just a question of negligence. That implies people knew but looked the other way. This was capture—ideological and philosophical capture. They were looking straight at [the problem] but didn’t see it. That has to do with a long process that started in the 1970s in the discipline of economics, which was captured by a group of people, often called the “Freshwater School”, based principally but not only in Chicago. They believed in rational behaviour and believed that the role of economists was to tell the state to get out of the way. They thought it was governments getting in the way that slowed down innovation and wealth-creation.
This group of economists gained influence over key economic journals, such as the American Economic Review and the Journal of Economic Literature, which determine the careers of academic economists around the world. So the whole profession got captured, and then generations of students. The subject became highly quantitative and more and more conceptual. Economics lost touch with reality. If you ask an average economics student to solve a common problem like unemployment, their answer would be no better than a taxi driver’s, because they don’t have the toolkit. But ask them to solve a simultaneous equation and they’d be very good at it! There was an obsession with data at the expense of ethics, intuition and judgement. This is a major issue—it reinforced the views of the Federal Reserve in the US, but also the Bank of England and others, regarding deregulation.
There are lots of scary aspects to the financial crisis, in terms of systemic risks, but what I find particularly troubling is that it is by far the best of the global systems. These people [in the Fed, the Bank of England and the Treasury] aren’t stupid, and they had the most extraordinary data. So you have to ask yourself why they were so blind [to the problems in the system]. Part of it was ideology. Part of it was being blinded by the blizzard of data. And part of it was politics. There was a short-termism and euphoria in politics and markets that was very difficult to puncture. No one had the guts to stop the credit boom in housing because it would be too costly politically. That is a structural problem. And it’s not a problem that’s been solved.
Why was the contagion in the global financial system so rapid in the late summer and autumn of 2008?
It was rapid because finance is at the frontier of globalisation. It’s probably the most globalised system. Most people in the developed world have some exposure to a pension fund. There’s virtually no one in the developed world who wasn’t affected by the financial crisis, and affected immediately. What was particularly significant was the creation of new financial instruments, particularly credit derivatives. In essence, derivatives are not bad things—they’re perfectly good and legal, and are designed to slice and dice risk and to sell it on. The problem is that when it’s on-sold a thousand times or more, you get a Ponzi scheme-type effect, where some pension-holder in London is holding a piece of liability which bears no relationship to the original source. And when the origination collapses, the whole thing falls down like a pack of cards. The complexity of the whole system was not properly mapped. The nodes were not understood in any significant way. The crisis also illustrates the vast gulf between this global system and national regulation.
So the moral you’d draw from the crisis is that the regulatory defence against such systemic crises has to be global?
Global and national. How global and what global institutions you need depends on the problem you’re trying to solve. If you’re trying to stop tax arbitrage in finance, then you want tiny places like the Cayman Islands, Lichtenstein, Luxembourg and Monaco to be part of your structure. But if you’re trying to solve credit derivative trading problems, then you want New York, Frankfurt and London to be involved. So it’s not that everything has to be equally global. You want the players that are going to affect the outcome to be part of an agreement. And before you can design that kind of variable architecture, you need to do the complex mapping. You need to understand where activity is happening and where the linkages are.
One of the consequences of that, as you point out in the book, is that often the regulation is as complex as the systems it is trying to fix.
This is the biggest challenge. You can’t fight complexity with complexity. That’s a recipe for disaster. You’ll get bogged down, as the Dodd-Frank Wall Street Reform and Consumer Protection Act in the US has got bogged down, in writing rules. The only people you’re going to satisfy are lawyers and accountants. Dodd-Frank has something like 33,000 pages of regulations. And they’ll be challenged. Lawyers will have a feast, and so will accountants. And it’s not going to stop the next financial crisis. Equally, the financial reporting dimensions of the act, which are hugely burdensome on US banks and corporations and which require them to report every transaction, are going to create an even bigger blizzard of data.
You need intuition and judgement—and Andrew Haldane, of the Bank of England, has written about this extremely eloquently in his article “The Dog and the Frisbee”. One of the things you need to be able to do is map who the significant players are. You need to think very carefully about what the likely source of risk is going to be to any given system. And we need to think carefully, too, about what the costs of regulation are relative to the potential benefits.
Let's now consider the question of inequality and other “social risks” that you discuss in the book. As you point out, one of the effects of globalisation has been a global convergence of average incomes, which has gone hand in hand with deepening of inequality within individual countries, particularly in the developed world.
Globalisation is such a powerful force for progress, in giving people opportunities to shape new ideas, to innovate, to produce and to sell. Those who are able to capture the benefits do very well. But those who aren’t able to engage, for whatever reason, fall further and further behind. Why can’t these people grab the benefits? They could be in a geographical location which is disconnected—think of certain parts of England or the Sahel, for example. They could have a skillset that’s not malleable. They could be illiterate or elderly. So how societies distribute the benefits of being connected to the rest of the world matters a great deal.
That’s inequality within countries. Between countries, it is the case that there’s a very rapid convergence, because emerging markets, on average, are growing at three, four or five times the rate of the old OECD economies. They are going to be rapidly converging. Indeed, there are parts of those countries that have already converged. For example the Shenzen district in China has higher average per capita incomes than Sheffield in the UK. Not all countries are growing, however. There’s a tail of about 20-25 countries—failed states, countries at war—which are not growing, places like Syria or Iraq, Zimbabwe and Sudan. These are falling further and further behind. If you look at the spread of all 202 countries in the world, there’s a divergence. But if you chop off the tail, there’s hyper-convergence. The third way to look at inequality is individuals across the world. There you find divergence as well. The share of the top one per cent of the world’s wealth is concentrating, as is the share of the top ten per cent. And that’s because of globalisation.
What’s so bad about inequality? Why are high levels of income and wealth inequality in a society undesirable?
Inequality is highly corrosive for society in a number of respects. I don’t think you can get a common project, a common view about which way your society is heading, and about the sacrifices you’re prepared to make, in your tax system and other things, unless you believe that you have a common outcome. So I think it’s extremely divisive for politics. The second reason I think it’s highly divisive is related to what Joseph Sitglitz writes about in his book The Price of Inequality. The problem is particularly acute in the US, but it’s also true in Brussels and the UK—and it’s that money buys power.
You’ve made a powerful argument against inequality on the grounds of justice, fairness and democracy. There’s another argument that some economists are making—even the IMF is now making it—and it’s that inequality is also bad for growth.
It’s delightful to see that the IMF is now saying things some of us thought we’d never see it say in our lifetimes! This is also a position now taken by the OECD. The argument is that inequality frustrates demand. It undermines recovery. If people don’t have jobs, they can’t buy things. So you can’t build effective demand with high levels of unemployment and low wages.
Ian Goldin and Mike Mariathasan's "The Butterfly Defect: How Globalization Creates Systemic Risks, and What to Do About It" is published by Princeton University Press (£25)
That unprecedented integration of economic and other systems has brought with it previously unimaginable benefits to substantial portions of humanity. But it also exposes us to "systemic risks" of a kind we've not faced before, and which politicians, at both the national and global level, need to try to understand—and quick. They could do worse than read "The Butterfly Defect," though its analysis of the risks that metastasise in fiendishly complex global systems will probably keep them awake at night.
I met Goldin in London earlier this week and began by asking him what he thinks the principal gains and benefits of globalisation have been.
IG: I think the last 25 years or so—the period since the fall of the Berlin Wall—have seen the most remarkable progress that the world has ever known in many dimensions of human wellbeing. Life expectancy, the reduction of poverty, quality of life, nutrition etc. If you look at this period in the long sweep of history, it will come to be seen as a period in which humanity leapt ahead in many dimensions. I attribute that largely to the coming down of walls and integration, which is how I define globalisation. It’s the flows between societies, principally the flow of ideas, that has lead to these changes.
Now, with that also comes great risk. For example, climate change, pollution, the destruction of biodiversity and so on are negative spillovers [of globalisation]. Then there is a series of new, emergent risks. I see the financial crisis of 2008 as one of the clearest exemplars of these new kinds of risk. It’s what I call the “butterfly defect"—small perturbations somewhere leading to these cascading, amplified shocks which can be deeply destabilising. Because there’s a mismatch between the integration processes and global governance, the risks are growing. You can’t manage globalisation at the local or national level alone. It needs management at those levels, and there’s a lot you can do, but you’re not going to solve climate change, pandemics, cyberterrorism or the problems of finance in one country. And that’s the tension.
JD: Central to your analysis of globalisation is the notion of “complexity”. Your claim is that complex systems incubate “systemic risks”.
Right. There’s a vast body literature that comes from physics, in the first instance, about how systems evolve and about how they can be made more robust and resilient. It’s also about how the nodes and networks in systems need to be understood in order to understand system stability. It is also important to understand attribution—cause and effect—in a system. And unless you model complex systems, it’s very difficult to know where the beginning and the end is. They become like entangled webs.
Now, part of the problem that politicians face is this attribution problem. They can no longer say [to voters], “I know the cause of your problems.” Now the problems they face come from beyond [national] borders. Bookshelves have been filled already with books about the financial crisis. But no one is going to be able to prove that it was caused by a, b or c. It’s too complex. Let’s take an example: if you think that systemic risk is likely to be propagated through key nodes—one bank, like Lehman Brothers, or one airport, like Heathrow, or one port, like Houston—you can design competition policy to ensure that no single place is too big to fail. If you think that systemic risk in transport hubs is a crucial factor in thinking about the stability of the airline network in the UK, for example, you might take a different view about the growth of Heathrow.
This isn’t the first age of globalisation, of course. But you do think technological progress and innovation makes the present age of globalisation distinctive, and that it makes the attendant risks distinctively threatening don't you?
Yes. This is a very different kind of globalisation to the previous waves that have gone before. This one is different for a number of reasons. One is the number of countries involved. Previous waves of globalisation tended to involve small sets of countries—for example, Britain globalising with its colonies. This wave is much broader because it’s across multiple dimensions. Virtually everything you can think of is being globalised. Then there’s the number of countries involved. There’s really only one country left in the world that’s not integrated [into the global system], and that’s North Korea. So whereas 20 or 30 years ago, half the world’s population was not connected and open, today maybe one per cent is.
The third important factor is the speed of integration, which is related to technology. It’s not only physical connectivity, it’s virtual connectivity. So that’s fundamentally different. Speed of transmission has changed.
You write that one of the features of the contemporary landscape is the unravelling of the “classical distinction between risk and uncertainty”. What do you mean by that?
Risk is something that you can quantify. It has a probability associated with it. Uncertainty is different: you cannot assign a probability to it. What I call “uncertainty” is not what Nicholas Taleb means when he talks about “black swan” events. Those are exogenous events. What I’m talking about are long-tail events. They’re different because they’re endogenous, they’re caused by us, by globalisation, and we could stop them happening if we understood complex systems well enough and governed them properly. One of the problems in unstable systems is not only how problems amplify and cascade from a single source, but also how they can hop between sectoral boundaries—how a pandemic, say, can lead to a financial crisis, or how a drought can lead to a war. I define “systemic risks” as risks that spill over borders. Borders between countries, but also borders between sectors.
Let’s turn now to the financial sector, specifically. I was struck by what you were just saying about endogenous risks and uncertainties. Is there an affinity between your argument here and the work of Hyman Minsky on financial instability?
I think there is. And also with the work of Keynes and Robert Shiller on herding behaviour, “animal spirits” and so on. These [financial] systems are inherently unstable and the market is unstable. They also tend to monopolistic power, because companies can pay for the lobbying which stops new entrants to the market. That’s why you need competition policy.
Among the many failures that led to the blow-up of the global financial system in 2008 was, you suggest in the book, a failure of the economics profession. You mentioned regulatory failures and failures of competition law just now. There was a fairly sustained assault on competition law, antitrust legislation and so on waged for many years by economists at the University of Chicago—one thinks of the work of Ronald Coase, Robert Bork and other members of the “Law and Economics” school.
Right. It’s not just a question of negligence. That implies people knew but looked the other way. This was capture—ideological and philosophical capture. They were looking straight at [the problem] but didn’t see it. That has to do with a long process that started in the 1970s in the discipline of economics, which was captured by a group of people, often called the “Freshwater School”, based principally but not only in Chicago. They believed in rational behaviour and believed that the role of economists was to tell the state to get out of the way. They thought it was governments getting in the way that slowed down innovation and wealth-creation.
This group of economists gained influence over key economic journals, such as the American Economic Review and the Journal of Economic Literature, which determine the careers of academic economists around the world. So the whole profession got captured, and then generations of students. The subject became highly quantitative and more and more conceptual. Economics lost touch with reality. If you ask an average economics student to solve a common problem like unemployment, their answer would be no better than a taxi driver’s, because they don’t have the toolkit. But ask them to solve a simultaneous equation and they’d be very good at it! There was an obsession with data at the expense of ethics, intuition and judgement. This is a major issue—it reinforced the views of the Federal Reserve in the US, but also the Bank of England and others, regarding deregulation.
There are lots of scary aspects to the financial crisis, in terms of systemic risks, but what I find particularly troubling is that it is by far the best of the global systems. These people [in the Fed, the Bank of England and the Treasury] aren’t stupid, and they had the most extraordinary data. So you have to ask yourself why they were so blind [to the problems in the system]. Part of it was ideology. Part of it was being blinded by the blizzard of data. And part of it was politics. There was a short-termism and euphoria in politics and markets that was very difficult to puncture. No one had the guts to stop the credit boom in housing because it would be too costly politically. That is a structural problem. And it’s not a problem that’s been solved.
Why was the contagion in the global financial system so rapid in the late summer and autumn of 2008?
It was rapid because finance is at the frontier of globalisation. It’s probably the most globalised system. Most people in the developed world have some exposure to a pension fund. There’s virtually no one in the developed world who wasn’t affected by the financial crisis, and affected immediately. What was particularly significant was the creation of new financial instruments, particularly credit derivatives. In essence, derivatives are not bad things—they’re perfectly good and legal, and are designed to slice and dice risk and to sell it on. The problem is that when it’s on-sold a thousand times or more, you get a Ponzi scheme-type effect, where some pension-holder in London is holding a piece of liability which bears no relationship to the original source. And when the origination collapses, the whole thing falls down like a pack of cards. The complexity of the whole system was not properly mapped. The nodes were not understood in any significant way. The crisis also illustrates the vast gulf between this global system and national regulation.
So the moral you’d draw from the crisis is that the regulatory defence against such systemic crises has to be global?
Global and national. How global and what global institutions you need depends on the problem you’re trying to solve. If you’re trying to stop tax arbitrage in finance, then you want tiny places like the Cayman Islands, Lichtenstein, Luxembourg and Monaco to be part of your structure. But if you’re trying to solve credit derivative trading problems, then you want New York, Frankfurt and London to be involved. So it’s not that everything has to be equally global. You want the players that are going to affect the outcome to be part of an agreement. And before you can design that kind of variable architecture, you need to do the complex mapping. You need to understand where activity is happening and where the linkages are.
One of the consequences of that, as you point out in the book, is that often the regulation is as complex as the systems it is trying to fix.
This is the biggest challenge. You can’t fight complexity with complexity. That’s a recipe for disaster. You’ll get bogged down, as the Dodd-Frank Wall Street Reform and Consumer Protection Act in the US has got bogged down, in writing rules. The only people you’re going to satisfy are lawyers and accountants. Dodd-Frank has something like 33,000 pages of regulations. And they’ll be challenged. Lawyers will have a feast, and so will accountants. And it’s not going to stop the next financial crisis. Equally, the financial reporting dimensions of the act, which are hugely burdensome on US banks and corporations and which require them to report every transaction, are going to create an even bigger blizzard of data.
You need intuition and judgement—and Andrew Haldane, of the Bank of England, has written about this extremely eloquently in his article “The Dog and the Frisbee”. One of the things you need to be able to do is map who the significant players are. You need to think very carefully about what the likely source of risk is going to be to any given system. And we need to think carefully, too, about what the costs of regulation are relative to the potential benefits.
Let's now consider the question of inequality and other “social risks” that you discuss in the book. As you point out, one of the effects of globalisation has been a global convergence of average incomes, which has gone hand in hand with deepening of inequality within individual countries, particularly in the developed world.
Globalisation is such a powerful force for progress, in giving people opportunities to shape new ideas, to innovate, to produce and to sell. Those who are able to capture the benefits do very well. But those who aren’t able to engage, for whatever reason, fall further and further behind. Why can’t these people grab the benefits? They could be in a geographical location which is disconnected—think of certain parts of England or the Sahel, for example. They could have a skillset that’s not malleable. They could be illiterate or elderly. So how societies distribute the benefits of being connected to the rest of the world matters a great deal.
That’s inequality within countries. Between countries, it is the case that there’s a very rapid convergence, because emerging markets, on average, are growing at three, four or five times the rate of the old OECD economies. They are going to be rapidly converging. Indeed, there are parts of those countries that have already converged. For example the Shenzen district in China has higher average per capita incomes than Sheffield in the UK. Not all countries are growing, however. There’s a tail of about 20-25 countries—failed states, countries at war—which are not growing, places like Syria or Iraq, Zimbabwe and Sudan. These are falling further and further behind. If you look at the spread of all 202 countries in the world, there’s a divergence. But if you chop off the tail, there’s hyper-convergence. The third way to look at inequality is individuals across the world. There you find divergence as well. The share of the top one per cent of the world’s wealth is concentrating, as is the share of the top ten per cent. And that’s because of globalisation.
What’s so bad about inequality? Why are high levels of income and wealth inequality in a society undesirable?
Inequality is highly corrosive for society in a number of respects. I don’t think you can get a common project, a common view about which way your society is heading, and about the sacrifices you’re prepared to make, in your tax system and other things, unless you believe that you have a common outcome. So I think it’s extremely divisive for politics. The second reason I think it’s highly divisive is related to what Joseph Sitglitz writes about in his book The Price of Inequality. The problem is particularly acute in the US, but it’s also true in Brussels and the UK—and it’s that money buys power.
You’ve made a powerful argument against inequality on the grounds of justice, fairness and democracy. There’s another argument that some economists are making—even the IMF is now making it—and it’s that inequality is also bad for growth.
It’s delightful to see that the IMF is now saying things some of us thought we’d never see it say in our lifetimes! This is also a position now taken by the OECD. The argument is that inequality frustrates demand. It undermines recovery. If people don’t have jobs, they can’t buy things. So you can’t build effective demand with high levels of unemployment and low wages.
Ian Goldin and Mike Mariathasan's "The Butterfly Defect: How Globalization Creates Systemic Risks, and What to Do About It" is published by Princeton University Press (£25)
