- The disconnect between the eurodollar curve and the 10-year U.S Treasury yield is puzzling.
- It might continue for a little while as both ends of the U.S yield curve react to different drivers.
- The disconnect is not significant enough to call for a sharp fall in U.S. Treasury yields.
The long end of the US Treasury curve is generally and naturally driven by the short end of the money market curve. Since the 2008 crisis, the correlation between 10-year U.S. Treasury yields and the slope of the Eurodollar (ED) curve has been close to 1. Interestingly enough, there have been only a few episodes of disconnect: mid-2009 and today. If the relationship had held over the past few months, 10-year yields would be closer to 3.1% than 2.7% (see chart below).
The following chart shows what is at stake. The ED curve and the news flow have disconnected sharply. The news flow-implied level for the ED12/ED1 calendar spread should be 1.3 (much closer to what the 10-year U.S. Treasury yield suggests).
It can be explained by the fact that the Fed is shunning the weather effect on the economy and considers that it is not "significant" (sic) enough to alter the course of tapering. Yet, to me, tapering is more related to the risk premium on the long end of Treasury curve than to the timing and aggressiveness of the zero-interest-rate-policy (ZIRP) exit. As a result, the current reading would suggest that the short end, not the long end of the US Treasury curve, is mispriced.
It is not that obvious though, as the relationship between the ED calendar spread and the PCE deflator remains unchanged: a lower inflation figure implies a steeper Eurodollar curve. The relationship is contemporaneous, which implies that if inflation is low today, then the slope is steep today too. The current reading would suggest no mispricing.
Interestingly enough, the steepness of the ED curve is not driven by the short end but by the distant tenors: Yellen sounding more hawkish, "dots" communication and the "6-month rule."
So while the short end is mostly driven by Fed speeches and assessment of the economy, the long end has digested the tapering fear and is mostly driven by macro factors, whose recent moves are not very bearish for U.S. Treasury prices: PCE deflator far below 2%, falling economic news flow. Add to this the improvement in the U.S. public deficit (and a stabilization of the public debt as a share of GDP), and the huge inflows of foreign purchases of Treasuries (chart below) - foreign purchases of U.S. Treasuries are currently more than twice higher than the U.S. external funding needs. This explains why the tapering fear has subsided.
My econometric model based on ISM, stock returns, long run inflation expectations and Fed Funds suggests that there is no particular mispricing for the long end.
What could make UST yields move sharply one way or another?
- A more dovish Fed, as I have shown before, would make the ED curve reconnect with long-term yield rather than push the latter much lower.
- A sharp retrenchment in US stock prices? The fact that U.S. Treasury bonds are neither rich nor cheap can be seen in my Fed Model chart below. Violent stock/bonds arbitrages generally take place when there is a significant disconnect in relative returns between stocks and Treasury bonds. The peculiar feature of the post-tapering era (false may announcement + implementation in early 2014) is that there is no signal of a major mispricing. The longevity of the positive returns of the S&P 500 might be a concern for many but, for the time being, both stocks and Treasuries are posting positive return, so the performance spread is far from overstretched.
- The confirmation in April/May suggests the slowdown was not related to the weather but was pervasive and genuine. This would undoubtedly drive the long rate lower and come along with a flatter money market curve.
Bottom line: the disconnect between the message sent by the short end of the yield curve and the current level of long-term Treasury yields may not call for a rapid adjustment on each tenor of the curve. Even though the Eurodollar curve has disconnected from the news flow, it is far from obvious which segment of the curve should adjust in the very short run.
I would rather lean on the segmentation side where the long end follows a mix a bullish factors such as a current soft patch for economic growth, strong foreign purchases and a lighter issuance program by the Treasury. On the other hand, the short end is driven by a single factor: Fed qualitative forward guidance.
This situation could linger for a short while (as it did in 2009), but if I had to draw a scenario I would clearly continue to lean toward higher 10-year yields (3.25%) at year end: US growth momentum, end of tapering (even if the impact on yield has never been obvious - see chart below) and a return into positive territory of the correlation between stocks and yields. Stay short TLO.