The last time I lent out EUR18 billion….

This whole Cyprus situation is crazy for more than the obvious reasons, it reminds me of the last time I lent out EUR18 billion. In 2007 I was a repo trader at a large “overleveraged” US bank. One of my roles was to invest the firm’s leftover cash on a daily basis. Every day and throughout each day, the treasury department would phone up to let me know how much cash to invest. The bank’s reserves had to be invested in only the safest assets – European government (since it was all Euro cash) bond repo. This is a loan collateralised by government bonds. I would lend cash out to other banks and get government bonds in return for a period of one day.

To manage risk, banks adhere to a leverage ratio. This means that they borrow against the value of the assets on their balance sheet, similar to a home buyer borrowing so many times the value of their deposit. When the value of the assets go up, the bank’s leverage goes down. Similarly if the assets fall in value, the leverage ratio goes up. But as a profit maximising business,  banks maintain a specific leverage ratio so when the value of the assets rise, they would borrow more against the higher value of their assets. But what this also means is if the value of the assets fall, the bank has to sell assets on its balance sheet to lower the leverage ratio. The cash generated from the sale of these assets generates cash which ends up with the bank’s treasury and in the short term, is invested in short term assets such as repo.

So in the first half of 2007, the asset backed and mortgage backed securities (ABS and MBS) markets started to fall, which sent the bank’s leverage ratio higher causing the bank to sell off some assets. Sharp falls in the price of assets meant bigger asset sales to bring the leverage ratio back in line. The sales generated cash and on one particular day it led to me having to invest EUR18 billion in European government bond repo or to put it another way, lending out EUR18 billion. This was more like parking cash on an overnight or tomnext (starting tomorrow for one day) basis than any sort of risky lending.

Now where have I seen that figure lately…oh yes, the GDP of Cyprus is EUR18 billion and just shy of the EUR18.3 billion bailout (EUR10bln from Troika plus EUR5.8bln from bank levy etc plus EUR2.5 bln Russian loan extension).

This shows at least two things. One how small Cyprus is in the grand scheme of things (although very important given the ramifications for the rest of the Eurozone) and two just how much liquidity and lending banks were involved in pre-financial crisis. 

For more on bank exposures and bank runs during the financial crisis read the below extract from my paper, “Fear and Uncertainty in the Money Markets: The Role of Asymmetric Information during the 2007-2009 Financial Crisis.”

See References at bottom for top papers on bank runs and liquidity squeezes (fairly relevant given events in Cyprus).

 1.      Laying the Foundations of a Liquidity Shock

Non-Inflationary Consistently Expansionary, or  N.I.C.E,[1] is how Mervyn King, Governor of the Bank of England, described the decade before the 2007-2009 financial crisis. Although the decade was littered with events such as the bursting of the dot com bubble, an increase in the awareness of the security threats following September 11 2001 and corporate scandals, such as Enron, the capitalist system of developed countries appeared, on the surface, to be stable. By 2007, stock markets had enjoyed a long period of growth, the new Euro Zone appeared to be a success and Britain had not had a recession since the early 1990s. It was a time of spend today, pay later as consumers enjoyed low interest rates while household wealth was boosted by a long period of house price growth.

Although, there were several doomsayers, notably Rajan (2005) and Roubini (2006), expectations of the future were optimistic. This was as true in the markets as it was on the high street and can be illustrated by Figure 1. The chart shows the difference between two benchmark rates for secured and unsecured lending over a three month period in the U.S., Europe and the U.K.


Figure 1: The Libor[2]/OIS[3] spreads in US and UK, Euribor/OIS spread for Euro Zone. Source: Bloomberg.

As can be seen from Figure 1, in the period before August 9 2007, the spread between these rates was extremely stable in the US, Euro Zone and the UK.

Although, Figure 1 illustrates the stable conditions of the interbank market before August 9 2007, it does not show the growing problems in the subprime mortgage market.

Over the previous ten years, the mortgage-backed securities (MBS) market had grown very quickly, initially encouraged by a policy by President Bill Clinton to address the growing poverty-wealth gap in the United States[4]. The aim was to make it easier for the poorer members of society to buy their own house by easing the borrowing constraints that existed for people with very little of their own capital. The intentions were good but this proved to have unintended consequences. The policy was continued by President, George W. Bush, and under his leadership, regulations were eased to allow banks to package together sub-prime mortgages with prime mortgages to create tradable assets similar to a bond. This process is called securitisation. Gorton (2008) explains the securitisation process, describing a market so complex, both in terms of how these products were structured and the network of investors involved, that it was impossible to know where the risks to investors and banks lay.

Gorton (2008) notes that in 2006, new synthetic indices for asset-backed securities were introduced to the market, known as ABX and this improved the publicly held information with respect to both the value and risk of MBS. Plus, it provided investors a tool with which to hedge exposures by buying and selling protection against falling prices. Gorton says that the common knowledge generated by the introduction these indices led ABX prices to fall due to the demand for protection, as well as the ability of investors to short the market putting even more downward pressure on prices. When house prices began to fall at the end of 2006, this is what happened.

These risks were either not overtly clear or disregarded by the broader market and it was assumed that only small players in the financial markets such as hedge funds would be lost. However, when the French bank, BNP Paribas, announced large losses in two of its hedge funds and stopped investor withdrawals on August 9 2007, the crisis escalated as the markets realised the problems were much larger than previously thought, with many of the world’s largest banks heavily exposed to MBS, the whole banking system was now threatened.

However, the reason for the escalation of the crisis wasn’t necessarily due to the absolute size of the MBS losses, as Brunnermeier (2009) writes, since the overall losses from the mortgage market were relatively modest compared to the $8 trillion of U.S. stock market wealth lost between the all-time high in October 2007 and October 2008. The reason for what has been described as the worst crisis since the Great Depression[5] is the result of an over-reliance by banks on short-term funding to finance longer-term assets.

The repo[6] market is one of the key tools banks use to manage their funding requirements. According to Hördahl and King (2008), repo markets doubled in size between 2002 and 2007 with roughly $10 trillion in both US and Euro-Zone repo markets and $1 trillion in the UK repo markets.


Figure 2: The growth and decline of overnight repo trading before and during the financial crisis. Source: Federal Reserve Bank of New York.

This form of financing was particularly popular with primary dealers especially the former investment banks Morgan Stanley, Goldman Sachs, Lehman Brothers, Merrill Lynch and Bear Stearns who did not have a customer base of depositors to form the basis of their capitalisation. This meant that due to their over-reliance on short-term funding, these banks were much more at risk to a liquidity freeze than other deposit taking banks.

These banks raised financing through the issuance of short-term debt by issuing commercial paper[7] as well as in the term repo market, while the overnight repo market was used to fine tune daily liquidity. Banks with a large customer base of depositors did not have to rely on this form of financing to such a degree and this is why the primary dealers were most at risk from the liquidity crisis that ensued. Figure 3 shows the effect this had on primary dealer credit default swap (CDS)[8] prices. The cost of insuring the primary dealers’ debt rose much higher than retail banks such as Bank of America and JP Morgan. This is particularly clear in the period after September 2008. However, it can also been seen that the market had concerns about Lehman brothers as far back as June 2008.

bank cds

Figure 3: The chart shows how CDS prices behaved between January 2007 and December 2010. Source: Markit.

Adrian and Shin (2010) explain how banks actively manage leverage[9] in order to maximise profits by using the assets on its balance sheet to finance other investments. Since the loan is a fixed size, if the asset price increases, the total leverage a bank is exposed to falls but if the asset price falls then leverage will increase. This is an example of passive leverage management. Since banks had fixed leverage targets set by the Value-at-Risk (VaR)[10] measure, however, they undertook active leverage management. This means that when the price of assets increases, instead of allowing the firm’s leverage to fall, the bank actively puts to use the spare capital that can be raised by maintaining a specific leverage ratio and one way to do this was in the overnight repo markets. However, if the assets’ value falls, the bank must sell assets to reduce leverage.

Several problems arose from this strategy. Adrian and Shin (2008) find that since banks generally all had a similar policy, the banking system was particularly vulnerable to a shock. For instance, if all banks held similar assets, there is a danger of creating a negative feedback loop when asset prices dropped because all banks would be selling the same or similar assets to reduce leverage at the same time. In turn, this would put even more downward pressure on asset prices, leverage would rise again and the same process of reducing the balance sheet would take place. Brunnermeier (2009) describes this as a liquidity spiral as banks are forced to sell assets but also face increasing haircuts.[11]

Liquidity spirals of this nature accelerated in March 2008 , when Bear Stearns was forced into a buyout by JP Morgan, and September 2008, when Lehman Brothers collapsed, because assess to funding in the markets was blocked.

If a bank is over-reliant on short-term loans, the lender can decide either not to renew the loan or shorten the maturity of the loan. The shorter the maturity, the more often the borrower must access the markets to secure funding. This is why these banks were particularly vulnerable to a liquidity squeeze in the interbank markets. If lenders suddenly decided to reduce lending and hoard cash, as happened from August 9 2007, banks would find it more difficult to secure funding.

Morris and Shin (2004) note that when a liquidity shock hits, traders’ trading horizon shrinks. In other words, the trader becomes much more concerned about securing funding for the immediate future i.e. overnight, and less concerned about funding for longer periods. This is because if funding cannot be found for today, then tomorrow does not matter since the trader will have lost his job.

Figure 4 shows the average interest rate on European general collateral.[12] At the key dates where banks were particularly stressed in August 2007 and September 2008, the repo rate spiked higher.


Figure 4: Chart shows the average daily rate for overnight European general collateral repo between 2005 and 2011. Source: Bloomberg.

Usually, primary dealers used the overnight market to invest excess funds that had been held back in reserve by the bank’s treasury department until the last moment in case of a shock or unforeseen funding requirement. But the spike in repo rates indicates that banks were relying on the overnight markets to raise funds because lending from other sources had dried up.

The collapse of liquidity drove Lehman Brothers to collapse on September 15 2008 while Merrill Lynch was saved by Bank of America. The two remaining Wall Street investment banks, Morgan Stanley and Goldman Sachs faced extreme pressure and were forced to become holding banks.[13]

A move against a bank such as this, represents asymmetric information, says Brunnermeier (2009). In this case, the market suspects a bank of struggling to finance its positions and rumours build in the market adding to the reluctance to lend to it. In the end it becomes a self-fulfilling prophesy. Since the primary dealers all financed themselves in the short-term markets, the liquidity freeze spread to the other associated banks leading to a systemic problem or bank run as Uhlig (2010) describes.

[1] Speech given by Mervyn King, Governer of the Bank of England October 14 2003.

[2] Libor = London Interbank Offered Rate. It is an unsecured interest rate since it is the average interest rate banks receive from interbank deposits. Libor is the daily reference rate for a range of maturities, from overnight to twelve months, at which banks can borrow unsecured funds from one another. The rate is the adjusted average of a group of designated contributor banks fixed at 11 am London time. Euribor is similar but it is fixed at 11 am Brussels time for the Euro currency.

[3] OIS = Overnight Index Swap also called EONIA swaps in the Euro Zone, SONIA swaps in the UK. An OIS is an agreement to swap a fixed rate of interest for a floating rate. The floating rate is the average of the overnight interest rate fixings (EONIA, SONIA, effective Fed Funds rate) over the whole period of the trade. At the end of the period, only the difference between the fixed rate and the average floating rate is exchanged between counterparties. No notional is exchanged; therefore, the OIS rate is equivalent to a secured interest rate.

[4] Rajan (2010).

[5] Crisis described by the IMF as the worst crisis since the Great Depression in April 2008

[6] A repo is a collateralised loan where the borrower pledges an asset as collateral in return for a loan. The presence of collateral makes the loan less risky for the lender because if the borrower defaults on his loan the lender still holds the asset. This arrangement means that interest rates charged are lower than other forms of loan such as unsecured loans.

[7] Commercial paper (CP) are short-term debt instruments, equivalent to US T-bills, but issued by a company and not the government.

[8] Credit default swaps (CDS) are used by investors to insure against the event of a loan default. For example, the seller of a Morgan Stanley CDS agrees to protect the buyer in the event that Morgan Stanley defaults on an underlying loan, such as Morgan Stanley issued commercial paper.

[9] Leverage – Banks pledge long term assets as collateral to gain access to funding. A bank’s leverage is the proportion of the balance sheet used to raise funding. It is defined as the ratio of total assets to equity, where equity is the difference between the value of the assets and the size of the loan used to buy the assets.

[10] Value-at-Risk (VaR) is a widely used probability measure that calculates the risk of loss on a portfolio of assets.

[11] A haircut is a discount on the value of a collateralised loan or repo. For instance, a 5% haircut will mean the borrower receives only $95 million while having to pledge $100 million of collateral in return. During the financial crisis, haircuts increased on risky assets as firms demanded greater protection from volatile prices.

[12] General collateral is the term used to describe a basket of government bonds used as security on a loan as part of a repo, as opposed to identifying specific bonds.

[13] Holding banks are forced to take on lower risk and are regulated more stringently, helping them to raise capital. Furthermore, as holding banks, Morgan Stanley and Goldman Sachs were required to increase their retail customer base of depositors adding to the banks’ income and capital base.


Adrian, T. S., Shin, H., S., (2010). Liquidity and Leverage. Journal of Financial Intermediation,Vol. 19, Iss. 3, pp. 418-437.

Brunnermeier, M. (2009). Deciphering the Liquidity and Credit Crunch 2007-2008. Journal of Economic Perspectives, Vol. 23, No. 1, pp 77-100.

Caballero, R., J. (2009b). A Global Perspective on the Great Financial Insurance Run: Causes, Consequences and Solutions. MIT mimeo, January 20.

Caballero, R., J. (2010). Macroeconomics After the Crisis: Time to Deal with the Pretence-of-Knowledge Syndrome. NBER Working Paper Series.

Caballero, R., J. (2010). Sudden Financial Arrest. IMF Economic Review, Vol. 58, No.1.

Caballero, R. J., Krishnamurthy, A. (2008). Collective Risk Management in a Flight to Quality Episode. The Journal of Finance, Vol. 63, No. 5.

Diamond, D. W., Dybvig, P. H. (1983) Bank Runs, Deposit Insurance, and Liquidity. Journal  of Political Economy, Vol. 91, No. 3, pp. 401-419.

Gorton, G., B. (2008). The Panic of 2007. NBER Working Paper Series.

Hördahl, P., & King, M. (2008). Development In The Repo Markets During The Financial Turmoil. BIS Quarterly Review. Bank of International Settlements, December.

Morris, S., Shin, H., S. (2004). Liquidity Black Holes. Review of Finance, Vol. 8, pp. 1-18.

Pritsker, M. (2010) Knightian Uncertainty and Interbank Lending. Unpublished paper, Board of Governors of the Federal Reserve.

Rajan, R. G. (2010). Fault Lines. Princeton University Press.

Roubini, N. (2006). Why Central Banks Should Burst Bubbles. International Finance, 9(1) pg 87-107.

Uhlig, H. (2010). A Model of a Systemic Bank Run. Journal of Monetary Economics. Vol. 57, pp 78-96.

This entry was posted in Bank runs, Banking and tagged , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s