crosslines ch banner

Newsfile: September 2012

Let's all play bank crash

We don’t have the definitive book of economic theory yet, even from the brilliant Joseph Stiglitz, but a number of new economists are chipping away at the crumbling foundations.

By Peter Hulm

Every game theorist and his graduate students, it seems, have been trying to bring down our financial system -- at least in simulation -- since 2008.

We've had the real crash. What the simulators have found is that our conventional theories don't predict what actually happened. Some economists, such as Joseph Stiglitz, told us our ideas were flawed before the sub-prime bubble burst. Others had their suspicions.

"Recent insights from the behavioural sciences, ranging from psychology to biology and anthropology, as well as from neuroscience and experimental economics, provide evidence that human behaviour largely differs from what economic theory assumes as the foundation of microbehaviour," said the organizers of a three-day symposium and one-day workshop at ETH Zurich on 11-14 September.

Dirk Helbing of ETHZ, for example, is convinced "we may need to thoroughly rethink the basic assumptions of macroeconomic and financial theory. Making minor modifications to the standard models to remove 'imperfections' may not be enough. The whole framework may need to be reconstructed."

But how to do it? And what should this new economic model look like? Hence the simulations, to try to build networks that operate like real life, financial crashes and all.

Changing economics as we know it

Behavioural theory, systems analysis and network simulation are the new tools that promise to change economics as we know it. Some 200 top specialists in these fields -- including Stiglitz -- gathered at the rather elegant but old ETHZ computer-sciences building for the meeting sponsored by the Geneva-based Latsis Foundation, and 150 of them presented papers.

Banks and financial institutions were a major concern of the modellers.

Xiaobing Feng of Shanghai Jiaotong University, SHIFT, China, and his colleagues put together a database of over 30,000 public and private overseas banks from around the world to identify the systemically important institutions.

Matteo Luciani of ECARES -- the Brussels Free University -- and his colleagues ranked financial institutions for the risk they faced and applied the measure to all firms in the Standard and Poor's 500 for 2003-2011.

Their simulations taught them how crises develop and what factors play a key role in system failures.

The money sink

Kimmo Soramäki of Financial Network Analytics, Spain, noted that when a bank is failing, any funds sent to it stay into the account until the bank resumes operations. In the system, such a bank continues to receive payments. It thus "traps some of the total liquidity in its account, becoming a sink".

"As a consequence other banks run short of liquidity and queues will build,[...] eventually causing payments to be delayed." Running a simulation to make up for the lack of real data publicly available, Soramäki and his co-author developed SinkRank, a measure that "not only accurately predicts the magnitude of disruption by a given bank in a payment system but also informs about which banks are most affected by the failure".

Andreas Krause from the University of Bath in the UK and a co-author triggered a simulated bank failure to estimate the likelihood of an individual bank failing. They concluded that balance sheets are of limited influence on the crash. "The probability of a bank failing depends on the characteristics of the network of interbank loans and the market structure," they reported.

What the crisis taught them

The analyses also confirmed the validity of some measures governments introduced to damp down the spread of failure. Luciani and co learned that "the systemic risk in the financial sector peaked in September 2008, but was greatly reduced by the introduction of TARP" (the Trouble Asset Relief Program in the U.S., which extended credit to investment institutions). In Europe systemic risk began to rise again in April 2010.

Breaking down the figures, the researchers showed that risk for insurance companies "continued to climb up to the rescue of AIG" while deposit-taking institutions "experienced generally declining systemic risk from late 2007, in line with [the] burst of the housing price bubble".

Comparing their results with other research, the authors showed that risk from capital shortfalls "remained at sustained levels" while risk from interconnectedness declined post-2008. The findings, they told the economic modellers, "show the importance of including the interconnectiveness of the financial sector with firms in the real economy" if policy-makers want to measure the amount of systemic risk that economies face.

Distress without default

Giovanni di Iasio of the Bank of Italy pointed out that "distress" could be propagated in a central bank in a network "even in the absence of defaults". His team is using a method they call DebtRank to "estimate the total potential loss to the financial system caused either by an initial default of a single institution or by a common shock to several institutions". It can also indicate which groups of institutions may be important to the survival of the system.

Even within the system, researchers are discovering that old ideas don't work.

Carsten Detken of the European Central Bank told the symposium ECB analysis of the crisis showed that market-price based indicators proved "useless" as early warnings of system risk. Structural indicators needed watching, such as global credit gaps.

Help the current theory ignores

The modellers also identified necessary steps which current theory ignores. Xiaobing Feng and co found geographical locale still helps determine the risk banks face, despite multinationalization. Their recommendation: combine policies to control systemic risk as well as individual vulnerability. "The current policy stance [...] emphasizes only the safety of the individual bank," they point out.

Similarly, Massimo Molinari of the University of Trento, Italy, challenged "one-size-fits all" capital-holding requirements for banks imposed to strengthen networks' resilience to shocks. In a computational laboratory he created an interbank network with three different strategies towards mergers and acquisitions in order to test resilience. He found the network structure and holdings rules help determine the extent of damage to the system.

"Capital requirements should be network-varying," he argued. "Flat capital requirements force an inefficient allocation of net worth within the system [making] the system less resilient, particularly by forcing large banks to hold too much capital while small institutions have less than they need.

"One clear implication is that the regulator [should] closely monitor the structure of the network and its evolution over time because policy on capital requirements is sensitive to it," he remarked.

Information contagion, money illusions

Co-Pierre Georg of UC3M and Oxford University and a co-author from the Banque de France pointed that "information contagion" is also a feature of systemic shocks, as news spreads. "This gives rise to feedback effects aggravating interbank market freezes, credit crunches (i.e. substantially reduced investment in real assets), fire sales, and interbank contagion."

With their simulation they have also been varying hypothetical investor preferences on risk, returns and liquidity to analyse the effect of proposed measures such as countercyclical capital requirements, raising liquidity ratios, setting a leverage ratio, and imposing surcharges on financial institutions.

Jean-Robert Tyran of the University of Vienna entitled his paper "The Economics of the Money Illusion" -- a reference to the tendency to think about economic transactions in terms of nominal money figures rather than real values. "While standard economics assumes that all economic agents are free from money illusion, increasing evidence suggests that thinking in nominal terms is common," he said. "Purely nominal changes can affect individual choices, and [...] money illusion can shape outcomes in labour, housing and asset markets."

Internal causes for price changes

Vladimir Fiilimonov of ETH Zurich and his co-author Didier Sornette quantified "the distance of the financial market to a critical state" depending on how much price changes are due to internal feedbacks as distinct from external news. They looked at the Chicago Mercantile Exchange and the futures contracts traded from 1998 to 2010 using the E-mini S&P 500. In 1998 70% of trades followed exogenous information. By 2007 they account for fewer than 30%.

Jean Charles Rochet of the Swiss Finance Institute and University of Zurich contested most strongly (in the panel debate) any claims that economic theory has broken down. But he was also toughest on bankers in his proposals for “taming” financial institutions. With co-author Xavier Freixas of UPF Barcelona he modeled a “Systemically Important Financial Institution (SIFI) that is too big (or too international) to fail”. They argued: “Without credible regulation and strong supervision, the shareholders of this institution might deliberately let its managers take excessive risk” (fill in the name of whichever institution happens to be in your newspaper’s scandal section today).  He proposed “a systemic tax […] to cover the costs of future crises and more importantly establishing a Systemic Risk Authority endowed with special resolution powers, including the control of bankers’ compensation packages during crisis periods”.


Dealing with homo economicus

The message of the new economics is that it does not make sense to continue treating a society's members as acting solely from rational grounds. Homo economicus  -- the God of conventional economists -- is dead, or often asleep. Roberto Weber of the University of Zurich noted: "Behaviour can change dramatically in ways unaccounted for by current theoretical models. This evidence highlights the need for improved behavioural theories of equilibrium selection [economic choices]."

Christian Zehnder of the University of Lausanne conducted a series of controlled laboratory experiments that suggest contracts between firms can shape the treatment they expect (entitlements) after a trade has taken place -- "in strong contrast to both standard economic theory and established behavioural models".

Arno Riedl, Maastricht University, Netherlands, added: "Economic experiments that ignore the power of partner choice in social interaction are likely doomed to produce misleading predictions for field behaviour and to give wrong guidance for theory development."

One of the key questions economists want to answer is how likely are people to save? Matthias Sutter of the University of Innsbruck, Austria, tested how teenagers made choices in an experiment and found he could predict their savings decisions because these correlated with health-related behaviour such as smoking.

Lorenz Goette from University of Lausanne and his co-authors used a large-scale field experiment to show that "groups with stronger community participation render their members generally more altruistic and trusting towards those who reciprocate favours." This variation on Hobbes's idea that we become "nasty, brutish and short" with each other when we see everyone else as a stranger also had a corollary that politicians could take to heart when trying to persuade their societies to approve more complicated measures that declaring war on Wall Street: "Increased community participation enhances the strategic sophistication of individuals."

Where’s the new paradigm?

Stiglitz urged a new paradigm, bringing social constraints more into the picture and introducing circuit-breakers to reduce the contagion of panic reactions in crises. But he was not claiming to outline a full new theory in Zurich to put economics on the right path and clearly did not think anyone in national or international institutions had the intellectual or legislative power to carry through such a programme (or even the obvious parts of conventional economic theory for dealing with crises). Maybe we will have to wait for the definitive book on Information Economics that will earn him his third Nobel Prize*.


* His work on information asymmetry in markets earned him a shared Nobel Memorial Prize in Economics. He was also a lead author for the Intergovernmental Panel on Climate Change, which received the Nobel Peace Prize in 2007.