To some extent, yields are low precisely because they are expected to rise. For example, five year gilts now yield 1.5 per cent while 10-year ones yield 2.6 per cent. Why, then, should anyone who has the choice prefer the five-year bond? The answer is that if you expect yields to rise, then you'll be able to reinvest the five-year gilt when it matures at a higher rate. If five-year yields in five-years' time are 4.1 per cent, then the total 10-year return on two five-year gilts will be equal to that on offer now on 10-year gilts. Simple maths thus tells us that the market expects yields to rise; this is the expectation hypothesis of the yield curve.
But there might be more than this. It’s possible that longer-dated yields are low because gilts and US Treasuries are over-reacting to low short-term rates, and so are under-estimating the longer-term dangers that bonds face. History gives us three reasons to believe this:
■ Longer yields are almost as volatile as shorter ones. In the US since 1985, the standard deviation of monthly changes in 10-year yields has been almost the same as that of changes in two-year yields.
■ Ten-year yields are very sensitive to moves in two-year ones. Since January 1985, each 0.1 percentage point monthly change in two-year yields has been associated with an average move of 0.082 percentage points in 10-year yields in the US, and with a move of 0.057 points in the UK. In the US, the correlation between the two has been 0.91; in the UK, it has been 0.73.
■ The correlation between the yield curve (measured by the gap between 10 and two-year yields) and subsequent changes in 10-year yields is slightly negative in the US and UK; this is true for one or three year changes. This is inconsistent with the expectations hypothesis, which says that an upward-sloping curve predicts rising long yields.
This evidence is difficult to reconcile with the idea that 10-year yields should be equal to the average of future two year yields, and so be less volatile than two-year yields, and less sensitive to changes in them. However, it is consistent with the idea that longer-term yields are excessively sensitive to short-term economic conditions, and so fall too far in bad economic times - when short rates are low - and rise too much in good times.
It's widely agreed that equities can be more volatile than they should be. Why shouldn't the same be true for bonds?
Now, all this does not prove that there is a bond bubble - though it's consistent with the possibility that there is. What it does mean is that the fact that yields are low now is not, in itself, a reason to dismiss the possible longer-term threats to bond markets.
