As trend followers, for the most part, the CTA industry promotes itself as having ‘insurance-like’ properties.
Specifically, this refers to performance during crisis periods like the bursting of the dotcom bubble in the early ‘noughties’, or the credit crisis around 2008.
During these periods equities1 were down 46% and 49% respectively while the Newedge CTA Index was up 33% and 14% respectively.
What is implicit in this typical analysis is that ‘crisis’ is often defined in equity terms. During the same two periods bonds2 returned 25% and 13% respectively: there was no crisis in fixed income.
This begs the question, how do CTAs perform during a fixed income crisis? This is not a question originating just from intellectual curiosity.
The Purple Book, which provides information on the UK pensions landscape, says that schemes now have larger exposure to fixed income than equity assets3.
So the performance of CTAs during fixed income sell-offs should be a concern to these managers. Further, with many Euro-zone government bonds now trading with sub-zero yields, the question has become quite a timely one.
The reason the answer is not widely known stems from the fact that there has not been a sell-off in fixed income of anything like the magnitudes seen in equity markets over the period that many CTA managers such as ourselves have track records, namely around two decades. Figure 1 shows this clearly through the strong clustering of fixed income returns to the right of the y axis, and few, if any, to the far left. To see a strong fixed-income sell-off we revisit the US market in the 60s and 70s.
We can simulate the performance of a CTA by running a simple momentum algorithm over a synthetic futures contract constructed using US Treasury yield curve data from the Federal Reserve. As we extend further back into the past, the range of rolling one-year periods exhibit a wider range of performance, in our case -20% to 30%. We plot our results by percentiles of these periods in Figure 2.
The results are intuitive. As expected, returns of a long-only strategy increase monotonically as percentile return increases.
The momentum strategy, on the other hand, shows the characteristic CTA ‘smile’ that is evident in Figure 1.
By choosing an extended timescale containing a range of fixed income environments, the characteristics of a fixed income momentum strategy are seen to be similar to those of an equity momentum strategy. This should come as no surprise; there is no long bias in the construction of a fixed income momentum signal.
A momentum strategy on any asset class can be profitable when price moves are strong and sustained in either direction, but it struggles when moves are mid-range.
In summary, the track records of many CTA managers are not long enough to have seen sustained fixed income sell-offs, but there is good reason to believe that CTAs can profit in this environment.
Source: Man Database. Rolling annual returns, sampled monthly, 1992-2014, a range typical of the longest CTA manager track records. Equity index represented by MSCI World Net Total Return Index hedged to USD, fixed income by Citigroup World Government Bong Index hedge to USD, scaled by three. Momentum returns from simulations.
Source: Man Database. Average annual returns, sampled monthly, 11963-2014. Momentum strategy looks at returns of a synthesized 10-year US Treasury future over the previous year. If this number is positive, the moment strategy goes long. If it is negative, it goes short.
1. MSCI World Net Total Return Index hedged to USD
2. Citigroup World Government Bond Index hedged to USD (Total return)