For fixed income fund managers it was once the case that if you understood the evolution of the relative sizes of the various cohorts of the young, the working, and the retired in a population, you could predict bond returns. Lots of workers relative to the “unproductive” young or elderly meant low wage pressures, lots of demand for savings assets such as bonds, and lower government borrowing. When workers became scarce, wages would rise, inflation would increase too, and bonds would sell off. The demographically based bond models worked a treat in the 1970s, 1980s and 1990s. But sometime around 2005 there was a complete breakdown in the relationship between bond yields and demographics. Based on the model alone we’d expect double digit long dated gilt and US Treasury yields, rather than the 1.6% and 2.4% we have today.