2011 Prospects for the US Swap Curve

I provide a rapid estimate of the current level of the US swap curve. My model points to an excessive steepness and we expect a 40/50 basis point tightening of the 2/10-year slope this year.

Gauging fair value

Before I start providing some insight on likely moves for the US swap curve in 2011, a glance at our fair value model may be useful. The first chart is drawn from a simple model of the 2-10 swap slope based on VIX, 10-year UST yield and the Eurodollar Futures calendar spread (ED1-ED9). The fit of the equation is correct with 95% of all observations falling into the +/-33 basis points’ bracket. The curve is too steep according to this metric (61 bp above the fitted level).

The second chart provides some interesting disconnection between the ED1-ED9 spread and the 2/10-y swap slope since mid-2009. Both curves move in sync but there has not been any spread reduction so far. It may be attributable to the implementation of quantitative easing policies.

To assess the likely impact of QE, we first chart below the ratio of the Fed balance sheet to both the total US government debt and notes & bonds outstanding. One clearly sees that in spite of huge bond purchases, the widening of the fiscal deficit brought the ratio sharply on the downside over the last few quarters.

Including this Fed/Supply factor into the model does not improve significantly the overall explanatory power but suggests a lower extent of excess steepness.

Excess steepness from where ?

The long end of the swap curve remains highly correlated to the 10-year UST yield (chart below). In addition, the 10-year swap spread, in spite of its historical directionality, remains in the 0/20 bp bracket. Therefore, the most likely explanation of the current shape of the 2/10 swap curve has to be found in the short end.

First, the implementation of the ZIRP policy broke the relationship between the swap curve and the Fed Funds target (second chart on the right). Second, for the first time since 1999 the spread between the 2-year swap rate and the Eurodollar calendar spread is excessively wide (first chart below). If history is a guide, some mean reversion should occur this year.

I estimated the statistical relationship between the 2-year swap rate ED1-ED9 spread. The results are visible on the right hand chart above. There is clearly a structural break in late 2009 as the residuals (orange line) drifted downward.

Mean reversion would suggest either a steeper ED1-ED9 (more Fed tightening expectations) or a lower 2-year interest rate – or, what happened in the past, both moves.

Implications for 2011

My main scenario is that the Fed will not hike before long, that is not until the start of 2013 (not in 2011 anyway). There is still too much slack in the economy and the tax-cut-prolongation impetuous will not last. Yet, uncertainties on the potential hawkishness of the FOMC and commodity driven inflation risk are mounting. As a result the long tenors of the Eurodollar curve will continue to embed a high risk premium. The ED1-ED9 slope is therefore close to be fairly priced.

The spread between the 2-year swap rate and the Fed fund rate is far above its medium run average. This suggests a potential of 40 basis points increase for the 2-year US swap rate. The first chart below is a monthly data model of the US swap 2-year rate. It clearly highlights the potential for higher short term yield. This same model also highlights the potential of flattening of the us swap curve - second chart on the right, based on fed funds (-), Vix (+), monetary base (-), public deficits (-) and S&P500 (+).

A clear risk for my forecast would be an upward breakout of the US 10-year swap spread. We compute below the major drivers of 10-year swap spreads and the risk associated:

1. Public bonds issuance will remain elevated and put a cap on further swap spread widening. The corporate/public issuance pattern remains favorable for tight swap spreads.

2. The directionality risk linked to higher UST yield is limited given the huge amounts of global liquidity, the benign core inflation backdrop and the downside risk on growth at year-end. This limits the convexity hedging risk (interest rates up=>duration up=> hedgers pay fixed=> swap rates go up).

3. Bank credit risk should not be ruled out but the 3-months Libor/ois spread is still 5 bp above the average pre-crisis level.

 

 

Wider 10-year swap spread seems unlikely in 2011, which confirm my call for a 40/50 bp flattening of the US 2/10-year swap curve.