Interactive Data Fixed Income Analytics-fixed income portfolio and credit risk analytics

The following article is reprinted from the March/April, 2002 issue
of On the Edge, the Interactive Data Fixed Income Analytics bimonthly newsletter.

Spread Volatility and Correlations

Kamel Bazizi, Vice President of Research & Consulting

The decrease in Treasury issuance since mid-1998 encouraged the growth in issuance of other fixed income products such as mortgage, corporate, and agency securities. Later that year, the Russian debt crisis and the near collapse of Long Term Capital Management resulted in wider and more volatile spreads. This shift to spread products and the increase in spread volatility motivated greater interest in measuring spread risk. The analysis of spreads is important for quantifying the spread risk of a portfolio versus a benchmark, often measured by the portfolio’s tracking error, which projects the standard deviation of the portfolio’s relative performance by comparing the portfolio and benchmark sensitivities to all market risk factors and considering their volatilities and correlations. One of the largest contributions to tracking error corresponds to a systematic widening or tightening of spreads. An analysis of spread volatility and correlation between mortgage, corporate, agency and swap spreads for the period 1994-2001, is presented in this article.

Over the past three years, fixed income markets have experienced higher spread level and greater volatility than at any other time in the past decade. The absolute level of spreads relative to Treasuries has also been much higher than in the past, and the correlation between spread products has been more significant. Spreads of mortgage, corporate, agency and swap securities have widened and tightened simultaneously while Treasuries moved around independently. Historical basis point spreads for these products are shown in Exhibit 1 for the 1/1/94 to 7/1/01 period. We can clearly see that, since mid 1998, spreads have been much higher and more volatile than in previous years. Exhibit 2 are some statistical measures presented to quantify the magnitude of these effects.

Exhibit 1: Historical Spreads

Mtg represents the 30-year FNMA current coupon yield spread to the 10-year Treasury
Fin represents the 10-year single A Finance yield spread to the 10-year Treasury
Agy represents the 10-year agency yield spread to the 10-year Treasury
Swap represents the 10-year Swap yield spread to the 10-year Treasury

Exhibit 2: Average, basis point volatility and correlation among selected spreads

The mortgage spread represents the difference between the 30-year conventional current coupon mortgage yield and the on-the-run 10-year Treasury rate. The corporate, agency, and swap spreads represent the yield difference between the 10-year single-A financial, the 10-year agency note, and the 10-year swap (respectively), and the 10-year Treasury.

In this analysis, year-by-year average spreads; standard deviations, and correlations of spreads were computed. Since the data was based on daily observations, standard deviations were annualized assuming 250 business days / annum. The percentage-spread changes were assumed to follow a normal distribution in computing standard deviations and correlations. The following illustrates how to interpret the results. In 2000, the current coupon mortgage yield represented an average spread of 178 basis points, with a standard deviation of 50 basis points.

It is interesting to look at the middle part of Exhibit 2, which consolidates year-by-year results into two time periods: 1/1/94-6/30/98 and 7/1/98-6/30/01. Clearly, the latter time period had higher average spreads and higher standard deviations of spreads than the earlier period. The latter period’s average spread for the 30-year current coupon conventional mortgage was 162 basis points (versus 109 basis points during the 1/1/94-6/30/98 period). The more recent period also had a standard deviation of mortgage spreads of 46 basis points (versus 28 basis points during 1/1/94-6/30/98). The reader may notice that the trends of other spread products are similar. For example, the 10-year swap spreads averaged 96 basis points over the last three years, versus 42 basis points during the previous period. The standard deviation of swap spreads during the latter period was 36 basis points, versus 14 basis points for the previous period. Spread products have also become more correlated than before. For example, the correlation of the corporate spread with the swap spread averaged 60% in the latter period versus 11% in the earlier period.

One reason for higher average spreads, greater standard deviation of spreads, and higher correlations among spread products, is that Treasuries have been increasingly scarce. The reduction in outstanding debt makes Treasuries more expensive than spread products resulting in higher spreads. The Russian debt crisis in 1998, the near collapse of Long Term capital started a new trend of even higher spreads and higher volatilities, as liquidity throughout the market declined and credit concerns increased. This analysis also showed that the spreads have become much more correlated than ever. The contribution of the spread volatility and correlation to a portfolio’s absolute and relative risk can be significant. For this reason, the volatility and correlation among spread products must be carefully estimated and incorporated into the portfolio risk management process.