Lehman should not be allowed to fall

I am about to mark the anniversary of the financial meltdown. In September 2008 bankruptcy of Lehman Brothers freeze the global financial system for several weeks. But the fall of Lehman was not the cause of the crises but rather a late symptom of cracks in the financial system caused by very high leverage. Financial intermediaries naturally leverage their balance sheets. But Chart 1 point to a strong positive relationship between changes in leverage and the size in balance sheet size which suggest a high procyclical in the leverage process of financial intermediaries.


Chart 1 reproduced from Hyun Song Shin

In good times leverage and balance sheet size of financial intermediaries increase together when risks decrease measured (VaR). But when the loss distribution is exponential, the behavior of intermediaries conforms to the Value-at-Risk rule, in which exposure is adjusted to maintain a constant probability of default. In a system context, increased risk reduces the debt capacity of the financial system as a whole, giving rise to amplified de-leveraging by institutions through the chain of repo transactions. This process leads to a synchronized contraction of balance sheets which will cause stresses that show up somewhere in the system.

Chart 2 US Fedwired interbank payment network reproduced from K.Soramäki

In case of the financial networks it is crucial for its sustainability to know where the stresses may show up as its topology is highly disassortative. I.E.US FED payment system contains over 9500 participating banks but the core of the network with 66 banks accounts for 75% of the daily transactions. Distribution of the connections between participants in that network follows the power law distributions (few banks are hubs for thousands of others while thousands of banks in that network are only connected to few others).

The good news is that this type of network (scale-free network) is very resilient to accidental failure as random failure even large amount of banks is unlikely to freeze the whole system ( the random failure of banks will take only the small ones as they are much more plentiful than the hubs ).

The Achilles heel of the network is its dependence on hubs (large interconnected banks) . Recent research suggests that generally speaking, that the simultaneous collapse of only of 5-15% of all hubs can crash free scale network. Therefore the protecting the hubs of the system are essential for its functioning. Allowing Lehman to collapse which was a central hub in financial network was clearly against this rule and was a major and cost full mistake.

Conclusions are rather fairly simple:

1) Regulators/central banker should watch and limit the leverage in the financial system (not only target inflation)

2) core of the network should be protected at all cost but also should be subject of stricter regulations