One of the main risks that Eurozone growth is facing is the high correlation between US (T-Note) and European (German, that is Bund) long-term interest rates. Even before the Fed mentioned the taper, European yields were above nominal growth, a situation that can easily lead to insolvency. Things might get worse as the threat of a taper leads to a new price range for US Treasuries.
The correlation between 10-year US Treasury and Bund yields remains stubbornly high, whether assessed in levels or in daily change. Here I propose a simple model of the spread and try to figure out to what extent a disconnect is possible.
My model is based on weekly data and a long run sample (1999-today). The T-note/Bund spread is driven by:
- The repo rate differential between the Fed and the ECB (positive sign)
- The VIX (negative sign): higher VIX generally spurs a search for safe haven assets and US yields tend to fall relatively more than their German counterparts
- The BTP (Italy)/Bund (Germany) reflects intra-EMU risks. It became statistically significant only in the late 2000s. Its sign is positive as a higher BTP/Bund spread gives a local status of safe haven to German bonds, not US Treasuries (the euro crisis spurred intra-euro capital flows, not extra EMU outflows)
- US monetary policy expectations: the T-note/Bund is driven positively by the spread between the 9th (≈2 years forward) and the first nearby Eurodollar contract
- The spread between US and Europe 2-year forward 3-month interest rate spread (proxy of expected monetary policy stance)
Interestingly enough, introducing the ratio of Fed vs. ECB would have given a much better fit but the positive sign is completely at odds with any theoretical foundation.
The model suggests that the current reading is above fair value and therefore that the spread has room to tighten. Really?
To better understand why the model suggests an overstretched spread, I provide the contribution of each input to the current level of the spread.
We can see that:
Contrary to 2008, 2010 and late 2011, the VIX level is neutral on the current reading of the spread;
- In spite of a significant retrenchment of the BTP/Bund spread and of the OMT insurance, the T-note/Bund remains structurally higher due to the special status of the Bund within the Eurozone
- Monetary policy expectations from the Eurodollar steepness have only had a limited contribution
- The sharp rise in the T-note/Bund spread has, unsurprisingly, been driven by the widening gap between expected repo rates in the US and in Europe (orange line).
Monetary policy expectations between the US and the Eurozone should diverge further in the future. Even if the Fed manages to properly explain its separation principal, the odds of investors pricing-in the first rate hike much later than Q4 2014 remains low. At the same time, the ECB will communicate on the "extended period of time" for many quarters for now, putting a lid on rate hike expectations.
The question is therefore: can the spread significantly rise above the limits reached in late 1999 and 2005? The answer is yes and the chart below explains why. I designed a "euro-neutral" BTP Bund spread, retrenching the BTP/Bund spread contribution to the actual level of the spread.
As can be seen above, the potential for the ex-euro-crisis T-note/Bund spread to rise is significant. If the slope of the Eurodollar curve steepens further and if, above all, the 2-year forward 3-month spread reaches its 2005 level, then the potential increase for the T-note/Bund stands close to 50/60 basis points (BPS) to a level only reached once in June 1999.
The differential between 10-year yields of US Treasuries and German Bund reflects a combination of traditional drivers (risk aversion, relative monetary policy expectations) and a structural change linked to the peculiar status of the Bund within the eurozone (safe haven).
My model shows that the euro crisis has added a (euro) risk premium to the T-note/Bund spread in the form of a structurally higher BTP/Bund spread. Therefore, the previous peaks of 2005 and 1999 should be disregarded as the traditional monetary drivers could push the spread well above the current reading of 100 bps.