Positive Linking and Financial Crises
This is the second of two contributed pieces by Paul Ormerod, the author of Positive Linking and, as I noted in my last post, in my opinion the most effective developer of multi-agent models of the economy.
Did Economists Go Mad? Networks and the Economic Crisis
The conduct of economic policy making over the ten to fifteen years prior to the financial crisis of 2008–9 exemplifies the fundamental problems of the conventional mindset of economics. At the time, it seemed as though clever policy makers devising clever rules and regulations to set the right incentives, to which economically rational agents would respond appropriately, had indeed solved key problems of macroeconomic management. Economic growth in the West was strong and steady, and both unemployment and inflation everywhere remained low.
Networks were conspicuous by their complete absence from the intellectual framework of policy makers. Yet network effects were absolutely central to the causes of the crisis.
There had been a number of scares along the way. In 1997–8, the rapidly growing area of East Asia experienced a financial crisis, with huge falls in output and employment throughout the region in 1998. But very soon growth and rising prosperity were restored. Network effects featured strongly in both the crash and the recovery. Doubts began to emerge about the economy of Thailand. In particular, there were worries – harbinger of bigger things to come ten years later – about whether the country was experiencing an unsustainable real-estate bubble. The Thai currency came under attack, and the government cut its link, its fixed value, to the US dollar. This was the signal for massive speculation against the currency and a sharp fall in its value.
But the crisis then spread like wildfire across almost every single country in the region. Even China, then nowhere near as connected to the rest of the world as it is now, suffered from a loss of foreign confidence. The financial collapse led to dramatic falls in output in many of the countries, with soaring unemployment.
Then, almost as suddenly, confidence returned, across both the networks of financial markets and the networks which create business confidence, or lack of it, in the domestic economies of the region. Why? Well, as the old saying goes, success has many fathers and failure is an orphan. The IMF was widely blamed for its role during the crisis, but claimed credit for restoring stability and paving the way for recovery. Even now, despite over a decade of the most intensive study by economists, we do not have an agreed answer to the questions why the economies collapsed so spectacularly but then, very quickly, bounced back as though nothing had happened. Networks have played little part in any of this analysis, but were undoubtedly the key.
So, a massive crisis in East Asia, the subsequent default on its debt by the Russian government, the collapse of Long Term Capital Management in America, the collapse of the dot.com bubble. Any single one of these might have triggered a worldwide crisis. But what might at any moment have turned into a bloodbath seemed to have been averted by the new-found skill and knowledge of policy makers, equipped not just with the longstanding tools of incentives but with the insights provided by the concept of ‘market failure’.
The intellectual underpinnings for the apparent miracle were provided by economic theory. Olivier Blanchard is the chief economist of the IMF. Here is what he had to say in August 2008 in an MIT working paper entitled ‘The State of Macro’: ‘For a long while after the explosion of macroeconomics in the 1970s, the field looked like a battlefield. Over time, however, largely because facts do not go away, a largely shared vision both of fluctuations and of methodology has emerged . . . The state of macro is good.’ The state of macro is good! In August 2008!
A few weeks later Lehmans went bankrupt. Capitalism itself was on the brink of another Great Depression on the scale of the 1930s when unemployment in the USA reached nearly 25 per cent. The period which had been dubbed the Great Moderation, when policy makers seemed to have been granted the touch of King Midas, in reality proved to be the Great Delusion.
Gradually, doubts began to seep across the network of banks, doubts about whether it was possible to know the true potential extent of losses of another bank to which money had been lent. Once banks became uncertain about whether they understood the true financial positions of other banks, they became reluctant to lend to each other.
Indeed, in August 2007 they simply stopped doing so, more or less completely. A real, no-holds-barred credit crunch.
The problem was not specific to any one bank, not specific to any incentives, any specific pricing of risk which had been undertaken. It was a network effect. A network effect which gripped most of the world’s banking system. One moment everything seemed fine and banks were happy to buy and sell these very complicated securitised bundles of loans. The next, almost in a twinkling of an eye, they were not. Indeed, they were extremely reluctant to carry out almost any sort of inter-bank trade, and specifically they stopped being willing to lend.
The price of each individual, isolated transaction had apparently been set optimally, the risk associated with it had been correctly assessed and taken into account. But banks had to believe that this was so. The belief had to be sustained across the networks on which the opinions and sentiments of bankers are formed. In Keynes’s phrase, the bankers were ‘a society of individuals each of whom is endeavouring to copy the others’. If they copied the opinion, the belief, that everything was fine, it would continue to be so.
But once they stopped believing, we had a credit crunch.
Banks, regulators, governments believed that the problem of pricing risk had been solved. Despite occasional doubts, occasional shocks, this belief persisted. Then, suddenly and dramatically, doubts about this spread like the Black Death across the networks of sentiment that run through financial institutions. And once this had happened, unlike Peter Pan exhorting the children to believe in fairies to save Tinkerbell from death, the authorities – regulators, governments, international institutions – found it impossible to exhort the banks to believe. Networks swamped all their efforts to restore confidence. Pessimism spread like wildfire.
Much publicity and controversy surrounded the setting up of the resulting Troubled Asset Relief Program (TARP), a $700-billion bail-out fund which required political approval and so was played out in full light of the democratic process in America. But in many ways this was of second-order importance to the purely administrative actions of the American authorities, who:
• nationalised the main mortgage companies, Fannie Mae and Freddie Mac;
• effectively nationalised the gigantic insurance company AIG;
• eliminated investment banks;
• forced mergers of giant retail banks; and
• guaranteed money-market funds.
The key point about all these actions is that the American authorities paid no attention to academic macroeconomic theory of the past thirty years. Real Business Cycle theory, Dynamic Stochastic General Equilibrium models, rational expectations – all the myriad erudite papers on these topics might just as well have never been written.
Instead, the authorities acted. They acted imperfectly, in conditions of huge uncertainty, drawing on the lessons of the 1930s and hoping that the mistakes of that period could be avoided. It was not a grand plan, nor did one ever exist. This was a process of people responding to events on the basis of imperfect knowledge and experimenting to discover what did and did not seem to work, desperately trying to restore confidence across financial networks. And networks were important to the outcomes, to the decisions which were made, at a very detailed level.
Confidence across networks was key. Confidence across networks of financial institutions that the monies owed to them by others would be paid. Confidence across networks of commercial companies that output was not about to collapse like it did in the 1930s, so that they would then not act in ways which made this a self-fulfilling prophecy. And confidence across networks of individuals that their worlds were not about to fall apart.
What they came up with worked. American GDP in 2009 fell by some 4 per cent compared to 2008, and by the autumn of 2011 the economy had not only stabilised but had grown for nine successive quarters. Indeed, by the third quarter of 2011, growth was sufficiently strong that the level of US GDP rose above its pre-crisis peak. (In contrast, between 1929 and 1930, the first year of the Great Depression, GDP fell by nearly 9 per cent, and the cumulative drop between 1929 and 1933 was 27 per cent.) Unemployment was still high, but employment had risen. The stabilisation programme worked, and a catastrophic collapse in output during 2009 was averted. It prevented pessimism and panic from percolating across networks.
It is a spectacular success of positive linking. The specific details of the measures which were taken were in general of second-order importance compared to the success in preventing the sentiment from spreading that America was about to suffer a re-run of the Great Depression of the 1930s.
The crisis in Europe during 2011/12 has mainly arisen through a failure of Eurozone governments to generate anywhere near the same degree of positive linking of sentiment across financial networks. We can usefully think of much of the economic policy in Europe as being not about specific measures in particular, the sort of thing which mainstream economists get excited about and whose impact they continue to believe they can measure, but about the desperate attempts by key policy makers to spread positive sentiment across the markets.
One thing is clear. Confidence is only weakly related to objective reality, to the actual facts. The principal concern is about public sector debt. In the case of Greece, the concern is entirely merited. Compared to the size of output in Greece (GDP), public sector debt is approaching 150 per cent, and there are few encouraging signs of a willingness to get to grips with the problem. With interest rates at around 7 per cent, this means that some 10 per cent ( 7 per cent of 150 per cent) of the total spending of the Greek government is going not on providing any form of services, but on paying the interest on its debt.
The comparable figure for Spain is only 60 per cent. Yet Spain, too, has experience repeated crises of confidence in the markets, and its interest rates have hovered around 7 per cent. In contrast, the interest rates on government debt in both the UK and Germany are not much more than 2 per cent, even though public sector debt is 80 per cent of GDP in the UK and nearly 85 per cent in Germany. Obviously public debt is not the single cause of the lack of confidence, but the Japanese currency is perceived of as being strong despite a public debt ration of nearly 200 per cent.
Policies which generate confidence are, once again, not so much the specific details, the economically ‘rational’ calculation of their potential consequences, but the creation of a positive mind set, a positive attitude on financial markets. Positive linking.