Low correlation doesn’t always mean better risk-reward
Imagine a strategy which captured all of the FTSE 100’s positive returns and none of its losses. This, surely, would be the ultimate investment – one generating only gains with no drawdowns. Sadly this instalment of Income Thinking doesn’t contain the secret of how to find this holy-grail. Instead we look at the correlation of this ‘ultimate investment’ to the FTSE. Surprisingly, the two would have been 84% correlated over the last 20 years.
Figure 1: The ultimate investment vs. FTSE 100
Source: Bloomberg as at 28.04.17.
A simple statistic
This is because correlation is a blunt measure which cannot differentiate between gains and losses. It tells you how, on average, an asset moves relative to another. It says nothing about how much it moves relative to the other asset and whether it captures more of the upside than the downside (called ‘skew’), or whether the relationship between these assets is likely to change at different points in the economic cycle. Yet the correlation of a fund with an investor’s domestic index is increasingly being used as a primary determinant of the suitability of an investment.
We believe, correlation needs to be viewed in combination with other characteristics to get a fuller appreciation of an investment’s behaviour.
Make hay whilst the sun shines
Throughout the majority of a typical economic cycle, growth-sensitive assets such as equities or high yield debt, have historically generated positive returns (for further detail, please refer to: Income Thinking: Structurally Diversifying Sources of Risk, November 2016). Over the last 50 years, approximately 60% of the monthly returns on the S&P 500 index have been positive. More striking is that outside of the overheating or recessionary stages of growth (when the cycle is either very mature or in outright contraction) approximately 70% of monthly returns are positive.
Figure 2: Average annualised excess returns of Growth assets at different stages of the business cycle
Source: Investec Asset Management calculations, GDP data from December 1969 to March 2016, Recession dates compiled by NBER, Compelling Forces scores start in October 1998, excess returns are the average of annualised monthly total returns of MSCI ACWI Index, FSTE EPRA NAREIT Developed index, BofA ML US High Yield Index, JP Morgan EMBI Global Diversified Composite Index, since inception to March 2016 in USD less the return on USD 1 Month Libor. The Compelling Forces Growth Score comprises three components: Fundamentals (recession indicators); Valuation (a yield-based approach to assessing asset class valuations); Market Price Behaviour (risk, momentum and economic surprise indicators).
It makes sense then to harvest the returns offered by growth-sensitive assets by having a high correlation to equities when the economic risks of doing so are low, and only adopting a more defensive position, or a low correlation to equities, as risks begin to rise. Achieving the best risk-adjusted returns by managing correlations appropriately through the business cycle should be more important than just targeting a low average correlation.
In addition to having a correlation that is suited to the economic environment, we believe the return profile of an investment can be improved by adjusting risk tactically to limit performance drawdowns. This is because, even in expansionary stages of growth, markets can be subject to losses caused by other drivers (such as geopolitics or sentiment).
As a result, a well-managed fund which looks to be positioned appropriately over both tactical and strategic horizons will vary its correlation to growth-sensitive assets over time. This can be seen more clearly by looking at correlation on a rolling basis rather than the simple long-run average which will necessarily miss any steps taken to mitigate losses.
Managing and evolving correlation
The effect of this can be seen in Figure 3 which shows the rolling correlation between the Investec Diversified Income Fund (DIF) and the FTSE 100. The chart illustrates the severe damping of correlations between the Fund and risk assets during the Brexit event and ‘risk-off’ episodes of 2015 and 2016. Similarly, when growth assets are felt to offer unattractive reward relative to their risks (for example, if valuations are lofty against a poor technical backdrop – such as in summer and winter 2015) then correlation can be dampened down for these ‘market-risk’ based reasons. Efficient methods for controlling correlation include the sale or purchase of bond and equity futures or the use of FX hedging. This allows the risk of the Fund to be controlled across multiple frontiers but without having to change the underlying positions and in so doing miss out on their income streams.
Figure 3: Rolling correlation DIF & FTSE 100 (90 days)
Source: Bloomberg as at 21.04.17.
A fuller picture of behaviour
The benefit of a correlation that varies over time can be an improvement in the skew of returns a strategy produces which is perhaps the most important measure of its attractiveness. Selecting resilient securities from the bottom-up, achieving structural diversification and tactically hedging against event and market risks, has led to lower drawdowns and better upside capture for the Investec Diversified Income Fund since launch. This allows better compounding of returns over time – a powerful attribute in a time of slim asset class returns.
Figure 4: Upside & downside capture of DIF vs. FTSE 100
Figure 5: Risk & return characteristics of DIF vs. FTSE 100 since inception
|DIF||FTSE 100 TOTAL RETURN|
|Compound annual return||5.4%||10.0%|
|% positive months||73%||66%|
|Annualised standard deviation||4.36%||10.29%|
|Sharpe ratio (0.0%)||1.22||0.98|
|DRAWDOWNS SINCE INCEPTION (PEAK : VALLEY) OF THE FTSE 100|
|Drawdown 1||30.04.13 : 30.08.13||(4.0%)||(12.1%)|
|Drawdown 2||31.08.16 : 30.11.16||(2.0%)||(5.3%)|
|Drawdown 3||31.03.15 : 30.04.15||(0.6%)||(3.7%)|
|Drawdown 4||28.11.14 : 31.12.14||(0.5%)||(3.5%)|
Past performance is not a reliable indicator of future results, losses may be made.
Source Figure 4: Investec Asset Management, in GBP gross of fees and taxes with income reinvested, UK Equity returns are for FTSE All Share Index, from 31.08.12 to 30.04.17. Investec Diversified Income Fund average monthly gain and loss as a proportion of UK Equities average gain and loss.
Source Figure 5: Bloomberg as at 30.04.17.
All information is as at 30.04.17 unless otherwise stated.
Investment objectives and performance targets will not necessarily be achieved and there is no guarantee that these investments will make profits; losses may be made.
Currency exchange: Changes in the relative values of different currencies may adversely affect the value of the Fund’s investments and any related income.
Default: There is a risk that the issuers of fixed income investments (e.g. bonds) may not be able to meet interest payments nor repay the money they have borrowed. The worse the credit quality of the issuer, the greater the risk of default and therefore investment loss.
Derivative counterparty: A counterparty to a derivative transaction may fail to meet its obligations to the Fund thereby leading to financial loss.
Derivatives: The use of derivatives may increase the overall risk in the Fund by multiplying the effect of both gains and losses. This may lead to large changes in the value of the Fund and potentially large financial loss.
Developing market: Some of the countries in which the Fund invests may have less developed legal, political, economic and/or other systems. These markets carry a higher risk of financial loss than those in countries generally regarded as being more developed.
Interest rate: The value of fixed income investments (e.g. bonds) tends to decrease when interest rates and/or inflation rises.
Multi-asset investment: The Fund is subject to possible financial losses in multiple markets and may underperform more focused funds.