The decline in bond yields over the past decades has supported profitability and diversification value of trend followers. Returns have been boosted by a persistent secular downward drift in interest rates and persistent positive carry. Diversification value owed to negative correlation of long duration positions with equity and credit returns, reflecting the dominance of deflationary over inflationary shocks. However, in a rising yield environment carry would work against the trend follower, while diversification value is dubious. A simple simulation that reverses the yield dynamics over the past 15 years suggests that a generic trend following strategy could perform poorly in a rising yield environment.
The post ties in with the lecture on systemic risks in the asset management industry on this site.
The past 30 years have been an era of predominantly falling interest rates. This secular trend has been one of the biggest drivers of positive performance of trend-following strategies, including those of Commodity Trading Advisors (CTAs). Persistently decreasing yields have not only boosted returns. Also, the predominant long exposure to duration has led trend following strategies to be well positioned for most of the large market setbacks of the past decades, including the emerging market currency crises of the 1990s, the burst of the dotcom bubble of the early 2000s, the global financial crisis of the late 2000s and the EU sovereign credit crisis of the early 2010s. Altogether, trend following strategies have proven to be an excellent provider of impressive diversifying strong historical returns for investors.
The success of trend following strategies has naturally led to asset growth at managers using them (see figure below). This rise in assets under management has forced the larger managers to take an arguably over-sized allocation to the biggest deepest markets, in particular the very same fixed income markets that have furnished so much of their historical returns. If rates continued to fall this would not be a bad thing for investors. However, as we have seen over the last 6 years, if yields are simply being kept at historical lows the performance of trend following strategies overall becomes less impressive.
Investors have long been impressed by the intricate level of strategy optimisation conducted to robustify returns, methods involving the likes of in-sample and out-of-sample or cross-validation that managers employ. However, some strategies have naturally been fitted to or biased by a facet of the historical data that may not be present in the future. Such strategies may soon find themselves broken. In particular, the fear is that interest rates may be rising in the medium term, raising questions as to how this will affect the performance of trend-followers and in particular their directional fixed income component.
In order to try to give an answer to this question and importantly to provide some unbiased colour to the discussion of how a bond trend-following strategy might perform in an era of rising rates, we have conducted a small experiment. To generate a rising yield environment the daily yield changes of the last 15 years were taken and then reversed going forward [figure below]. This regime was chosen simplistically so that there can be no accusation of choosing the data to suit any particular argument.
The below chart shows the total return series of the Schatz, Bobl and Bund futures historically and beyond the inflection point.
The next figure below shows the performance of a generic trend-following strategy for a position of equally-weighted Bund and US 10yr Treasury futures. For the past it has been very much in line with those which will be found in the largest, most successful CTAs. This experiment suggests that the generic medium-term strategy would produce sustained losses in the simplistic rising yield environment. This result and its implications are robust to reasonable changes in speed of signals and the allocation between them (while remaining medium-term) and sensible response functions.
In order to illustrate why generic trend following in a rising yield environment makes losses, the figure below shows the cumulative historical performance of the same trend-following strategy on the original, historical data set. This exhibit separates the data into returns when the strategy is long and short the traded markets.
Clearly, as one would expect given the historical environment, the strategy spends less time being short than being long and it is a good job too for investors given the performance. The performance of medium-term trend-following strategies when applied to the majority of markets presents the same picture. The general problem with going short risk is carry. Approximately 50% of historical returns in fixed income have come from carry. Typically, carry has been in the same direction as the price trend; going against it for any period of time is diffcult as it provides what amounts to a substantial drag. When yields rise, instead of being in line with the direction of the yield-induced prices, carry would in general be in the opposite direction.
It’s not all doom and gloom as there will be times when short-term moves dominate and should a trend follower be positioned appropriately then there are profits to be made. Although, in general, anyone who has swam in the sea knows how difficult it is to swim against the tide, and this, to some extent, is to be the challenge for trend-following CTA and global macro managers and their investors going forward. Perhaps, even more of a challenge is to ensure that managers that have previously provided downside protection will do so in the event of a raising rate environment.