Riskpremiums

A popular explanation for low bond yields is the decline of risk premiums. But how can risk premiums be accurately calculated? Christel Rendu de Lint of Pictet Asset Management looks at one method developed by a researcher at the Fed

One reason put forward by both the Fed chairman Alan Greenspan and the ECB president Jean-Claude Trichet to try to explain persistently low bond yields is that risk premiums have decreased. Unfortunately, verifying this statement is not easy as risk premiums are not clearly defined, let alone measured.

But Fed researcher Brian Sack has developed an interesting way to derive the risk premium or, more precisely, the term premium embedded in three-month Eurodollar futures contracts. His aim is to 'clean out' Eurodollar contracts of this term premium in order to obtain the market's true monetary policy expectations.

Sack calculates the term premium by comparing Eurodollar futures contracts four and five years out (using the same month of reference) and assuming that the pricing difference captures a risk premium. The rationale is that differences in pricing so far out in time cannot reasonably represent expected changes in monetary policy. Firstly it is hardly credible that market participants have a view so far into the future. Secondly the Eurodollar futures term structure is almost always sloping upwards, implying that investors always expect the Fed to tighten further. Sack therefore concludes that the difference between the implied three-month Libor rate in five as against four years must be accounted for by a term premium.

Sack's approach appears to deliver a plausible time series of an interest rate risk premium over the last decade. Why? Intuitively, we would expect such a risk premium to be linked to economic uncertainty and market volatility. We find that this is the case. As a proxy for economic uncertainty we use a two-year rolling standard deviation of the monthly change in the US Conference Board leading indicator and measure market volatility with the two-year rolling standard deviation of the two-year bond yield. We find that the above calculated risk premium is closely correlated with economic uncertainty and market volatility, lending credence to its derivation method.

As evident in chart 1, while the risk premium has declined sharply since mid-2003, it did so from an elevated level. On this measure, the risk premium is not lower than in the 1990s and, as such, will not help explain the low level of interest rates.

To be sure, we constructed a sort of risk-free 10-year interest rate by subtracting the above calculated risk premium (multiplied by 10) from the 10-year government bond yield. We then compared it with the annual growth rate of nominal GDP to see whether the resulting series looks plausible. We found that it does, with this risk-free 10-year interest rate roughly equal to nominal GDP growth from 1993-2001 (see chart 2). But over the last three and a half years, the risk-free rate was considerably lower than economic growth. In other words, Greenspan and Trichet's risk premium argument does not help explain low interest rates, rather the opposite (as the risk premium had in fact risen sharply as yields started declining in 2000).

The beauty, of course, is that risk premiums can be defined and derived in many different ways, and it is undoubtedly possible to find a measure that will support the above assertion. Yet, in our view, the culprit lies elsewhere. So back to the drawing board to explain low yields!