Financial Crisis and the Global Banking Network


The greek sovereign debt crisis has been going on for nearly a decade and while it’s not making headlines in the news anymore one can safely assume that it had an impact on the european banks. For instance, in 2011 european banks (excluding greek banks) held greek government bonds worth 45.8bn Euros and by 2015 european banks have reduced their exposure to 2.4bn Euros. It also highlighted the importance of interconnectedness of financial markets in the transmission of financial shocksas european countries tried to safeguard their own financial markets from the fallout of greek default. Hence, the question is how much the sovereign debt crisis affected the interconnectedness of financial markets.

From the interconnectedness of financial systems the risk of contagion from balance sheet effects arises. For instance, consider a bank in country A has lent money to a bank in country B. Now, the bank in country goes bankrupt and cannot repay the loan to the bank in country A. The bank in country A has now to write down the loans made to the bank in country B. Depending on how much it has lent to the bank in country B it may have to default as well and so on.

Despite the dangers of contagion in interconnected financial systems a banking network may also decrease the risk of financial crisis as diversification improves. In the previous example the bank in country A has lent money to the bank in country B. Now, instead of lending most of its money to one bank it diversifies its loan portfolio and lends to a lot of banks. Now, the bankruptcy of one bank poses less of a threat to its existence than before.

The Network

To study the interconnectedness of the european banking network we use the consolidated banking statistics of 53, of which 28 belong to the European Union, countries provided by the Bank of International systems. The consolidated banking statistics consolidate gross claims of international banking groups. These claims are then aggregated following the nationality of parent banks. For example, the french subsidiary of a german bank lends money to a spanish bank. Under the residency principle of locational banking statistics this would be counted as a loan from France to Spain while under the Consolidated Banking Statistics this would be counted as a loan from Germany to Spain as the risk would ultimately be borne by the german bank.

As countries with a large GDP and better developed financial systems have the ability to borrow and lend more it is important to normalize the data. Here, we chose to normalize the data by dividing liabilities towards a country by all liabilities. As all liabilities are greater than zero, this number would be between 0 and 1 with a 0 implying no loans are made towards a country and a 1 all loans are made towards a country. Hence, we call this measure ‘Exposure’.

The following figure shows a heatmap and on the x-axis we have the reporting country and on the y-axis we have the residence of the counterparty.

Figure 1: Interconnectedness of financial sectors in 2007 Q1 (Left Panel) and 2018 Q2 (Right Panel) as a heatmap.

Some takeaways:

  • 10 years after the great financial crisis the US financial system the US financial system remains important as most countries are heavily exposed to the US financial system
  • Despite the looming ‘hard’ Brexit, the UK still has the most important financial market in Europe.
  • While exposure to Greece was already small in 2008, it is even smaller in 2018.
  • Sweden is the Gateway to Northern Europe and became even more important in this role in 2018.

From the heatmap it is very difficult to observe which countries are important or not. Hence, we display the interconnectedness in a network graph. To make the network graph easier to read and interpret we say that two countries are connected if the exposure between these two countries exceed 10%. For example if a bank in country A has given a loan to country B and it exceeds 10% of the total loan portfolio we say that these two countries. While the cutoff-value of 10% is arbitrary it gives us some sense which countries have important financial connections.

Figure 2: Interconnectedness of financial sectors in 2007 Q1 (Left Panel) and 2018 Q2 (Right Panel) displayed as a network. The cutoff between connected and unconnected is 0.1. Yellow circles are EU members.


Financial derivatives, such as credit derivatives, played a crucial role in the transmission of financial shocks during great financial crisis. This transmission of financial shocks happened through balance sheet effects, i.e. after Lehman Brothers declared bankruptcy in September 2008 many counterparties had to write off their claims against Lehman Brothers. Furthermore, due to opacity of loan portfolios underlying credit derivatives these derivatives came under special scrutiny and their market value declined which put banks under even more pressure. And for the network graph we take the same cut-off value as before.

The next two figures show how the interconnectedness of financial markets via derivatives evolved. As it can be seen in 2007 the market was even more concenctrated for derivatives than for debt instruments with the most important countries being the United States, France and the UK. Exposure for derivatives is similarly defined as the exposure for loans in the previous sections, i.e. we divide the claims from derivative contracts between banks in country A and banks in country B by the total claims arising from financial derivatives.

This situation has changed in 2018 and while the previous three countries reamin important the importance of Germany and Switzerland has drastically increased. Thus, in many countries the financial sector has diversified the derivatives portfolio to reduce exposure and improve the managing of risks.