An important development in the interest rate swap market in recent years has been widespread use of collateralization to mitigate counterparty credit risk. When the market started in the 1980s, most swap contracts were unsecured and any imbalance in the credit standings between the two counterparties was “priced into” the fixed rate or managed by having the weaker party get some type of credit enhancement. In the 1990s, after the introduction of the CSA (Credit Support Annex) to the standard ISDA (International Swap and Derivatives Association) master agreement, posting cash collateral when the market value of the swap is negative became more common and nowadays is the industry norm.
The main implication of collateralization is that the credit risk on the swap transaction becomes minimal, similar to exchange-traded futures contracts. To be sure, futures require initial margin accounts by both counterparties to provide an additional buffer to potential default loss. Bilateral CSAs usually entail a zero threshold, meaning only the counterparty having the negative market value for the swap posts collateral, so there still is some “tail” risk of default. In any case, minimal credit risk on the swap implies that the discount factors to get the present value of the annuity for the difference between the contractual and current swap market fixed rates, which remains unambiguous, should be based on risk-free interest rates or, at least, nearly risk-free from the perspective of the rating agencies.
Why then is it not market practice to use actively traded U.S. Treasury notes and bonds, for which there are ample price data, to get the discount factors to value collateralized swaps denominated in U.S. dollars? The problem is that U.S. Treasury yields are in general too low for this purpose. Treasuries are by far the most liquid debt security in the fixed-income market and are in high demand as collateral for the repo transactions. Exemption from state and local income taxes lowers their yields even more. Also, Treasury yields typically are more volatile than swap rates because they are the first asset class to absorb fluctuations in demand and supply arising from international capital flows, especially during flights to quality.
The ideal discount factors to value collateralized derivatives contracts would come from traded securities having the same liquidity, tax status, and volatility as interest rate swaps but credit risk approaching zero. Pre-2007, fixed rates on LIBOR swaps were viewed to be a reasonable and workable proxy for the risk-free yield curve. However, in the post-2007 world the presence of a persistent and sizable spreads between LIBOR and overnight indexed swap (OIS) rates exposes the “credit risk approaching zero” presumption.
An overnight indexed swap is a derivative contract on the total return of a reference rate that is compounded daily over a set time period. In the U.S. dollar market, the reference rate is the effective federal funds rate. It is calculated and reported by the Federal Reserve each day in its H.15 Report and is the weighted average of brokered trades between banks for overnight ownership of deposits at the Fed (i.e., bank reserves). The effective fed funds rate is not necessarily equal to the target rate set by the Federal Open Market Committee (FOMC) and announced at regularly scheduled FOMC meetings. The Fed merely aims to keep the effective rate close to its target via open market operations of buying and selling securities. [In the Euro-zone, the OIS reference rate is EONIA (Euro Overnight Index Average), which essentially is the 1-day interbank rate. In the U.K., the reference rate is SONIA (Sterling Overnight Index Average).]
Until August 2007, the LIBOR-OIS spread was consistently narrow, typically just 8 to 10 basis points. That justified the use of rates on LIBOR swaps as proxies for risk-free transactions. Some commentators date the onset of the financial crisis at August 9, 2007, which was the day when the LIBOR-OIS spread first spiked upward. It remained high, oscillating between 50 and 100 basis points, and then jumped again in the fall of 2008, reaching its pinnacle at about 350 basis points after the announcement of the Lehman bankruptcy on September 15, 2008. It then returned to more normal levels in 2009 only to go up again in 2011 reflecting concerns over the Euro-zone sovereign debt crisis. The LIBOR-OIS spread has become a widely used indicator for bank credit and liquidity risk.
The OIS curve is now preferred by swap dealers because it removes the bank credit and liquidity risk that is being priced into LIBOR. Moreover, central clearing of standardized swaps and collateralization of uncleared transactions are mandated for dealers by the Dodd-Frank Act of 2010. In response, central clearers such as CME and LCH specifically use the OIS curve to value swaps and to establish collateral requirements.