Chart-01 shows national benchmark bond return indices for the five countries, and the equal-weight M5 basket, rebased to 18 months ago. The sharp drop in yield boosted returns on 10-year US-Treasuries, while two of the markets with negative yields (Switzerland and Japan) lagged the other three. Yields appear to have bottomed there. The odd one out is the German bond market: Yields in Germany declined further in already negative territory. British ‚Gilts’ seem to take their lead from across the Atlantic, not the Channel.
The patterns seen during 18 months in Chart-01 are also valid over 36 months: US and British bonds performed significantly better than their peers, with Japanese and Swiss bonds being the two poorest performing bond markets. This order is slightly modified when risk is taken into account, but US bonds remain on top, being the only bond market to have generated positive risk adjusted returns.
Chart-02 visualises the data from Table-01 in a scatter diagram (to the left) which shows normalised risk horizontally, and normalised return vertically. The dashed diagonal marks equilibrium. The scatter to the right shows identical data 12 months earlier. Bond markets have enjoyed yet another injection of anti-aging drugs (quantitative easing) but the response is rather differentiated, with Switzerland and Japan barely, or no longer, able to show a response.
The plot of yields across ten years helps to explain why bond markets are beginning to diverge, as all preceding illustrations have suggested. There is a limit to how far yields can drop and some markets, such as Switzerland, Japan, and arguably also Germany (with a number of additional European bond markets) have clearly exhausted that space. The US and Great Britain are much later in that process, but even these markets are seeing yields at record low levels regardless of length of historic context.
Naturally, the performance of bond markets (return indices) as shown in Chart-04 is the logical result of the history of their yields.
Using trails of normalised data (here 3-years) makes changes to indices more transparent. Chart-05 illustrates how bond markets had effectively come to the end of an already most unusually extended cycle (effectively in place since 1982 !) some time during 2018 and 2019. One can only hope that investors are aware of how sensitive bond prices are when coupons, or current yields, are no longer able to cushion against a reversal in long-term interest rates.
Chart-06 plots normalised risk (observed risk) over 3-year trails, just as Chart-05 did for normalised returnsExhibit-08 (below). Considering the mathematical base for bond prices as almost fully dependent on yield levels, an ex-post risk metric may not have the same meaning as it does for other asset classes, and forward looking risk is easier to quantify, at least in conjunction with yield forecasts. The readings in the chart can almost only move towards higher levels of risk, if only because the ‚coupons’ element of return is diminishing.
Table-02 gives values of all bond return indices at fixed points in time: ten, five, and three years ago, as well as 12 months back, latest year and and latest month end. In the lower portion of the table the corresponding return is indicates, emphasising the best (green) and worst (red) market in each time frame. Note the dominance of US returns for any time frame between year-to-date and five years ago, all due to performance during the most recent 12 months.
Chart-07 plots only the M5 basket, across the last ten years, re-basing the value to 100 at the starting point of the graph. In the lower portion, the erosion from peak value is shown. The scales have been set to be identical with those in an equivalent illustration for equities (found in a dedicated section).
Returning to numerical information, Table-03 investigates similarities, or differences across the five national bond markets during the latest ten years. The value changes shown in the upper portion of the table could already be gauged from previous illustrations.
Interesting are the results of running regression analyses of each national market and the equal-weight composite (as independent variable).Note surprisingly low correlations: no single market has r-squared greater than 50% and German ‚bunds’ show a higher correlation than the US market.
Chart-08 gives annualised returns calculated from 31.12.1999 to date. Over these more than 20 years of history, Japan and Switzerland are the clear laggards, two markets with traditionally low interest rates, and among the first to enter the twilight of negative bond yields. Even so, returns on Swiss bonds (the second worst investment in this universe) were still double those from Japanese bonds.