In March 2000, just as the NASDAQ index was hitting its peak, I began work as an academic consultant to a small group of staff at the Bank of England charged with formulating conceptual frameworks for thinking about financial stability, led by Andy Haldane and with Prasanna Gai as a key member. The world seemed more tranquil back then and financial crises seemed to be afflictions of emerging economies with poor governance, although the events of the summer of 1998 associated with the LTCM crisis gave us a glimpse that financial turmoil did not respect the neat taxonomy of developed and emerging economies. But the prevailing sentiment back then was that financial crises were exceptional, worthy of study in the same way that doctors study a rare disease, but something that remained properly in the laboratory rather than something that might impinge on the day to day decisions of economic policymakers.
Our group at the Bank of England ploughed on, nevertheless. We noticed early on that the traditional ‘domino’ models of contagion fell far short of sounding adequate alarm bells against potential fragility. The domino model works through cascading defaults whereby, if Bank A has borrowed from Bank B, while Bank B has borrowed from Bank C, and so on, then a shock to Bank A’s assets that leads to default will hit Bank B, and if the hit is big enough, Bank B’s solvency will be impaired, in which case Bank C would be hit, and so on. The trouble was that such ‘domino’ models almost never generate systemic crises, as balance sheet interconnections in real life are rarely large enough in practice to topple banks at the second or third round of defaults. These models gave a false sense of security to policymakers and suggested that systemic crises were remote possibilities that would not leave the laboratory.
However, the blind spot in the domino model was that banks are assumed passive observers, and banks further down the chain stand idly by while banks start toppling further up the chain. More realistically, Bank C would be running long before Bank A’s collapse leads to losses for Bank B. It was the run on the bank, not the hit on the solvency of the bank that would lead to the most acute systemic problems. This is a lesson that was driven home repeatedly in the recent crisis. Nor did the domino models take account of valuation (p.viii) effects where falling prices interacted with marked-to-market constraints to set off the ‘death spiral’ of leveraged institutions that became such a familiar feature of the recent crisis. Although we were aware of the potential impact of these channels, we did not quite realize how potent they would prove to be once the monetary policy frameworks at advanced economy central banks that neglected financial stability concerns allowed leverage to build up beyond breaking point.
The ideas developed by Prasanna Gai and his collaborators trace their origins to this early work by the team at the Bank of England. They show the richness of the conceptual frameworks for thinking about systemic risk and contagion. They draw on the microeconomics of banking, quantitative risk management, coordination games, and the theory of networks.
The monograph culminates with a description of the quantitative risk assessment framework developed at the Bank of England—the Risk Assessment Model for Systemic Institutions (RAMSI)—that incorporated the insights of the early research of the Bank of England team. RAMSI drew on those early lessons by integrating balance sheet-based models of banks with a network model in a way that allows for feedback effect of asset sales. Shocks and scenarios from a macroeconomic model are then fed through the framework to describe how risk profiles evolve through banks’ business operations. Many of the more recent models of systemic risk being developed at central banks draw on the insights from this early research.
There is still a long way to go in incorporating systemic risk fully into the policymaker’s toolkit, but this monograph take us a long way down the road we need to travel.
Hyun Song Shin
Hughes-Rogers Professor of Economics
4 July 2012