Exclusive: A Review of CTA Performance in 2016

March, 2017

Tom Wrobel, Director of Alternative Investments Consulting in SG's Prime Services division, today published his review of CTA performance in 2016.


Managed Futures strategies experienced an up-and-down year in 2016 and, despite positive performance in December, finished the year in negative territory on average. Conditions for trend-following strategies were challenging, and the strategy experienced both significant run-up and run-down periods as developing trends broke twice with no strong individual theme able to develop. It wasn’t all doom and gloom, however, as non-trend strategies, in particular short-term CTAs, were able to eke out a positive return.

This snapshot analyses the performance of managed futures strategies during 2016 using the SG CTA Indices, attribution data from the SG Trend Indicator, and simulations designed to show the market environment and opportunity set for momentum trading across varying time horizons and markets.


The Short-Term Traders Index, was our only index to post a positive return of our indices for the year, albeit only just up 0.41%. The SG CTA Index experienced its first negative annual return since 2012, finishing the year down -2.89%, as shown in Exhibit 1. Positive returns from non-trend strategies were undermined by the returns of the trend-following strategies, which lost -6.21% during the year as shown by the Trend Index. The CTA Mutual Fund Index, which completed its first year of live performance and has a higher proportion of trend followers, also underperformed the CTA Index by -2.56%.

Exhibit 1Index Summary Statistics for 2016


2016 Return

2016 Volatility

2016 Max Drawdown

SG CTA Index




SG Trend Index




SG CTA Mutual Fund Index




SG Short-Term Traders Index




SG Trend Indicator




S&P 500 TR




Source: SG CIB & Bloomberg

All of the CTA Indices' annualised volatilities were closely in line with their historical averages but experienced larger maximum drawdowns in 2016 than in recent years using daily data. The drawdowns were, however, contained in the order of 1.4–1.6 times of each of the indices’ volatilities.

The CTA Index constituents displayed slightly higher return dispersion than over previous years, as shown in Exhibit 2, in which the best and worst performers are at the ends of the whiskers, and the interquartile range is shown by the box. Although the overall range of returns in 2016 of 34.61% was within the historical norm, the interquartile range was slightly higher than the average: 14.33% compared to 12.53%. This was primarily the result of a particularly low third quartile equal to -9.40%, which matches the two previous lowest values, in 2009 and 2011. Compared to these years, however, 2016 displayed a slightly higher first quartile and a significantly higher maximum value of 19.69%.

Exhibit 2Return Dispersion for the SG CTA Index Constituents by Year

Source: SG CIB

The range of individual manager fortunes is detailed in Exhibit 3, plotting their returns and volatilities for 2016. Whilst the Trend Index fell by -6.21%, there are in fact two distinct clusters of managers: nine posted negative returns with an average of -9.82%, whilst three bucked the trend by producing positive returns, the maximum being 11.55%. The performance of the non-trend and short-term managers was much more evenly split between gains and losses, with the largest outliers both coming from short-term strategies. The overall range of returns was equal to 53.14%, between the maximum of 27.60% and lowest of -25.53%.

Exhibit 32016 Performance for SG CTA Indices Constituents

Source: SG CIB

CTA returns in 2016 saw four identifiable run-up and run-down periods as shown in Exhibit 4, which details the monthly returns throughout the year. The CTA Index started 2016 by gaining 7.27% with the strongest combined January and February returns since 2013 before giving it all back over the following three months to be flat on the year at the end of May. Political uncertainty would shape the remainder of the year, with the UK Brexit vote kicking off another strong run of performance in June and July. Uncertainty in the buildup to the US presidential election saw markets being more volatile and range bound, lacking clear direction, which resulted in a second period of significant losses for all CTA indices, with the CTA index entering negative territory for the year. CTAs closed the year with positive returns in December, but only the Short-Term Traders got back in the black.

Exhibit 42016 SG CTA Indices Monthly Returns

Source: SG CIB

A more detailed look at the returns of short-term managers over the year reveals a significant outperformance during the first drawdown period and following the US election as more responsive strategies were better able to adapt to the changing market environment, as will be shown later, and perhaps explains the outperformance of short-term strategies at the index level.


As in previous CTA performance reviews we calculate a Signal-to-Noise Ratio (SNR) for each of the 55 markets in our Trend Indicator portfolio. Ideal environments for trend following of strong directional moves with little volatility are represented by high SNR values. In contrast, volatile markets with limited directionality, or "choppy" conditions, would be more challenging and have a lower SNR.

To represent medium- to long-term trend following, we have combined the SNR calculations across markets and time frames to focus on look-backs of between 80–120 days. The combined SNR over the last five years is shown in Exhibit 5. The average SNR in 2016 was 4.7, which is relatively low; this level is comparable to 2012 and 2013, when CTAs experienced a prolonged drawdown.

The evolution of the SNR during 2016 sheds some light on the four periods of returns highlighted above, with a high SNR value of 6.2 in February that was the highest since early 2015, corresponding to CTAs' positive start to the year. Markets became significantly less “trendy”; the SNR was below 4.0 for much of March to May. Despite a small pickup in SNR on the back of the Brexit vote, uncertainty around the US election resulted in the lowest values for the year in November.

Market “trendiness” returned after the election result, with a similar increase in SNR to that seen at the beginning of the year, to reach an intrayear high of 6.5. Whilst December was a positive month for CTA strategies, it is interesting to note that the Index returns were not as strong as earlier in the year and the Trend Indicator outperformed the Trend Index by 3.42%. An analysis of the Trend Index constituent returns during December showed that the managers' return volatilities was, on average, 17.5% below where it had been during the year. We believe CTA managers may have been more dynamic in reacting to the more recent market environment than the Trend Indicator, which rebalances on a monthly basis and uses a look-back period of two years to size positions.

Exhibit 5Combined Portfolio SNR for 55 Markets (80–120 days)

Source: SG CIB & Bloomberg

We can evaluate the SNR values at a sector level to provide insight of the driving factors during the challenging periods of CTA performance. Exhibit 6 shows the average quarterly SNR calculations of each sector over 2016 and confirms that for the majority of sectors, the SNR values started at relatively high levels but fell gradually during the year, before increasing slightly again at the end of the year. The exception to this was the Equity sector, which increased in Q2 as trends in equity markets started to develop. None of the sectors indicate a consistently good environment for trend following, with a particularly low SNR value in Q2 in Bond markets, as directional moves that presented good opportunities for trend following were subject to sudden and volatile changes.

Exhibit 6Average 2016 Combined Sector SNR (80–120 days) by Quarter

Source: SG CIB & Bloomberg


The SG Trend Indicator lost -0.19% in 2016, placing it in the middle of the pack of the individual Trend Index constituents shown in Exhibit 3. The Indicator showed an outperformance of the Trend Index by 6.02%, though it should be noted that the performance for the Trend Indicator was exactly the same as for the Trend Index year-to-date on the 11th November. After this point, as we explained earlier, the Trend Indicator returns were more volatile than those of the constituent managers. This higher volatility, combined with trending market conditions, saw the Indicator returning +8.60% in the last few weeks of the year vs. 1.86% for the Trend Index. The Trend Indicator’s correlation to the Trend Index in 2016 was 0.69, slightly lower than previous years.

Exhibit 7 shows the return contribution by sector for the Trend Indicator split between H1 and the full year 2016. Overall, the Trend Indicator finished the first half of the year slightly positive, gaining 2.49%. The Bond sector was the main performance driver for the model in H1, contributing 7.48%. In contrast, there were no clear Currency trends, which saw this sector losing -2.91%, and Equity markets fluctuated, adding to losses.

The picture by the end of 2016 was significantly different. The emergence of trends in the second half of the year in Equity Indices and Currencies resulted in H2 gains of 5.18% and 4.34%, respectively, reversing earlier losses to end the year with positive contributions from both sectors. Bonds also slightly extended their positive run of performance, but whereas H1 gains were from long trends early in the year and after the Brexit result, H2 performance was primarily the result of short positions as bonds fell.

Exhibit 72016 SG Trend Indicator Performance Attribution by Sector H1 and Full Year

Source: SG CIB & Bloomberg

In contrast, the Commodity sector lost -9.00% in the second half of 2016, and every subsector ended the year with a negative contribution. The Energy complex was the biggest loser, impacting the portfolio -7.27% over the full year as a result of consistent losses from February onwards as 2015’s bear market came to an end, and our model was whipsawed long and short a number of times in a range-bound environment.

With the exception of energy markets, returns within each sector were relatively mixed, and Exhibit 8 shows the performance attribution by individual market, with evidence of strong trends in some markets being offset by more challenging markets. Energy markets experienced some of the worst negative returns; not just Crude Oil, which lost -4.50%, but also Heating Oil and Natural Gas suffered. Other commodity market results were more mixed. Although the Bond sector was the best performing sector, the US 2-Year bond proved more challenging, losing -1.00%. All other bond markets contributed positive performance, in particular the longer maturity 10- and 30-year bonds. Equity indices also had relatively mixed fortunes but ended 2016 with two-thirds of markets in positive territory, benefitting from the bull market in the latter part of 2016 after a frustrating start.

Exhibit 82016 SG Trend Indicator Performance Attribution by Market

Source: SG CIB & Bloomberg

The second-largest loser was Mexican Peso, which lost -2.75% after the model initiated a Long position vs. the US Dollar in November just before the US election result, which triggered more than a 12% fall to record lows. In contrast, the Currency sector also contained the largest individual positive contribution: Japanese Yen gained 4.20%. The model did a good job of capturing the sustained Long trend vs. the US Dollar, switching to a Long position on the first day of the year, as well as timing the equally large down move in Q4 of 2016. All sectors, with the exception of Energy, finished the year with market return dispersion, suggesting that no strong individual theme dominated the portfolio, and Exhibit 9 details the Portfolio Position Concentration (PPC) over the last five years. The average PPC value for 2016 was 0.54, in line with the historical average of 0.57, and there were three occasions when PPC peaked to almost 0.8. Not only are these peaks significantly lower than the high of 0.97 at the beginning of 2015, but they are short-lived. A PPC value of 0.8 indicates the development of a theme within the portfolio and a sector, with potentially a large number of markets positioned in the same direction, but the themes did not last long as the values quickly fell back down and markets become more independent.

The changing market environment that produced 2016’s inconsistent returns is seen in Exhibit 10, which counts the average number of days in position for the Trend Indicator over all markets and approximates to the average length of trends in the portfolio. There are two periods when trends emerge and the average position duration was higher than 100 days, and the buildup to these peaks correspond to the two positive runs of performance early on and midway through 2016. They are followed by sharp decreases as the established trends break and markets reverse. The lowest number of days in position occurred at the end of the year. In December, the average trade length went as low as 58 but then started to increase as new trends started to emerge and the monthly return was positive.

Exhibit 9SG Trend Indicator Portfolio Position Concentration

Source: SG CIB & Bloomberg

Exhibit 10SG Trend Indicator Average Number of Days in Position

Source: SG CIB & Bloomberg


As with previous snapshots, we will use a series of simple price-momentum models in order to look at the return opportunities across model time frames. Appendix 3 details the calculation methodology of these momentum models and the construction of the resulting heat map, which is shown in Exhibit 11.

The average z-score value in 2016 across all of the 4,400 model/ market combinations was -0.07, marginally lower than 2015’s value of 0.02. The Bond sector had the highest values, averaging 0.70 across parameters, and there are standout "warm" vertical strips in German bonds, which represent strong return opportunities from just one Long trade. The "coolest" part of the heat map is the top left, in the Commodity and Equity sectors with long look-backs, and much of the Currency sector. In these sectors, look-backs of greater than one year had an average z-score of -0.58, and these markets would therefore have been challenging for very long-term trend-following strategies. The exceptions were British Pound and Wheat, which both exhibit positive values over higher model parameters.

Within the Interest Rate sector the average z-score value was 0.02, and Short Sterling had a high average value of 1.57. Despite the potential for return opportunities, as in previous years the low volatility in these markets resulted in target position sizes that would have been too large relative to the available liquidity. The Trend Indicator’s portfolio-construction methodology is designed to prevent this from happening, and so position size caps were in place in all Interest Rate markets. There were also caps in place in 13 out of 17 markets in the Commodity sector, with risk reallocated to markets that can accommodate larger position sizes. The two most liquid commodity markets that can handle larger positions are Crude Oil and Gold, both of which had negative average z-scores and therefore contributed inflated losses to the portfolio. Scanning across the heat map, medium-term parameters were in general "warmer." The highest individual look-back was 135 days, with an average z-score of 0.34. We have approximated look-backs of 80 to 120 days to medium- to long-term trend following, for which the average value was positive: 0.08. Despite the increase in the average value, it is still approximately zero, so it is not surprising that our Trend Indicator posted a year-end return of -0.19%, with inconsistent returns across sectors and markets.


Managed Futures strategies experienced an up and down year in 2016, and only short-term trading strategies managed to eke out a positive return for the year. Conditions for trend following were mixed, with varying individual manager fortunes. Using a series of Signal-to-Noise Ratios, we examined how market conditions were unfavourable for trend following. After a strong start to 2016, political uncertainty shaped the remainder of the year.

Attribution data from the SG Trend Indicator showed that the two big impacts on returns in 2016 came from positive contributions from Bond markets and negative performance from the Energy complex. Despite the evidence of strong trends in some markets, there were two periods when developing trends broke, and no strong individual theme was able to develop.

Finally, we used a set of simple price-momentum models to look at returns across parameter sets in a heat map, highlighting the opportunities available in the Bond sector and certain individual markets. Overall, longer-term look-backs presented decreased return opportunities in 2016, and longer-term strategies may have struggled compared to more medium-term parameters.

Exhibit 11Heat Map of Model Time Frame / Market z-scores for 2016


Source: SG CIB & Bloomberg

Appendix 1

SNR Calculation

The n-day SNR for a given market is calculated by taking the absolute price change over an n-day period and dividing it by the average n-day volatility. We chose to use an average true range as the volatility measure as it is best able to capture more of the intraday and overnight volatility in a way that a measurement of Close-to-Close price change does not.

The SNR for a given market is therefore calculated as follows:

The output of this calculation is relatively intuitive, for example, an SNR value of five indicates that the market has moved five times the volatility (average true range) over the given look back period.

Average SNR values can be taken across multiple markets to calculate a Portfolio SNR. A Combined Portfolio SNR is calculated by averaging the Portfolio SNR across multiple values of n.

Appendix 2

Portfolio Position Concentration Calculation

Using daily positions history from the Newedge trend indicator we assign a position to every market. Each long position is given a ‘+1’, and each short a ‘-1’. For each sector (Commodities, Equity Indices, Currencies, Bonds and Interest Rates), we calculate the degree for position concentration as follows:

These values are then aggregated into the Portfolio Position Concentration  (PPC) as a function of sector weighting as below, this gives a range bound indicator between 0 – 1:

Source Paper: "SG Prime Services AlternativeEdge snapshot: How many trades are there currently?”

Appendix 3

Momentum Model Simulations Calculation Methodology and Heat Map Setup

Momentum Model and z-score Calculation:

In this paper we calculate a series of simple price-momentum models in order to look at performance across parameter sets.

These models are calculated using the following rules:

If Closing Price >= Closing Price n days ago, then go long at tomorrow’s close

If Closing Price < Closing Price n days ago, then go short at tomorrow’s close

We apply the above rules for values of n from 5 to 400 in intervals of five days, employing the continually adjusted futures series we use for the Newedge Trend Indicator. This gives us a total of 4,400 model / market combinations.

In order to show the return opportunity set available for these combinations we calculate a z-score (or Sharpe ratio) as follows:

z-score = [Model P&L using 1 futures contract] / [Standard Error of P&L]

Heat Map Construction:

The 55 markets are listed on the horizontal axis and are separated in asset class groups. The vertical axis represents the various look-back values (values of n) and is separated to show where a three-month, six-month, and one-year look back would sit.

Each of the 4,400 z-scores that we calculate is represented by one box in the heat map with the colours being determined the specific z-score. Squares that are coloured

p   Yellow or Red  Illustrates a good opportunity set for that model / market combination

p   Green                Illustrates flat performance

p   Blue                   Highlight negative returns

In this heat map we have assumed frictionless markets and zero transaction costs in order to seek out pockets of heat where there may have been an opportunity set. For those models with short look-back periods the impact of these transaction costs would have been significant, and what may appear to be a good model / market combination may not prove to be profitable when actually traded (this is very dependent on the execution skill of a manager).

You can download the pdf files here - choose the loation relevant to you.





Any views and commentary in this communication are short-term views of Prime Services, a division of Société Générale Global Markets within Société Générale (“SG”).
This communication has not been produced by the Société Générale Research Department and is not Investment Research, Non-Independent Research, Research Recommendations, personal recommendations or any other form of recommendation.
Any views and commentary in this communication are short-term views (together, “Views”) of our Sales teams and/or Trading desks from which it originates (the “Authors”). This communication has not been produced, reviewed or approved by the Société Générale Research Department, and is not subject to any prohibition on dealing ahead of its dissemination. The Views are not objective or independent of the interests of the Authors or other Société Générale Group Sales and/or Trading personnel, who are active participants in the markets, investments or strategies referred to in this communication. The Views are not a personal recommendation and do not take into account whether any product or transaction is suitable for any particular investor.
SG, its affiliates and associated personnel may at any time acquire, hold or dispose of long or short positions (including hedging and trading positions) which may impact the performance of a product.
This communication is provided for information purposes only and it is subject to change. This communication is indicative only and is not binding.
SG is not offering to sell or seeking offers to buy any product or enter into any transaction. Any transaction requires SG’s subsequent formal agreement which will be subject to internal approvals and binding transaction documents.
SG is not responsible for the use made of this communication other than the purpose for which it is intended, except to the extent this would be prohibited by law or regulation.
SG is not an advisor and will not provide any advice relating to a product. Before making an investment decision, investors should ensure they have sufficient information to ascertain the legal, financial, tax and regulatory consequences of an investment to enable them to make an informed investment decision.
SG is not responsible for information stated to be obtained or derived from third party sources or statistical services. Any past or simulated past performance (including back-testing) contained herein is no indication as to future performance. Estimates of future performance are based on assumptions that may not be realised. Investments in general and derivatives in particular involve numerous risk, including amongst others, market, counterparty default and liquidity risk. Trading options or futures involves additional risks and is not suitable for all investors. An option may become worthless by its expiration date, as it is a depreciating asset. Option ownership could result in significant loss or gain, especially for options unhedged positions. Prior to buying or selling an option, investors must review the “Characteristics and Risks of Standardised Options” available at http://www.optionsclearing.com/about/publications/character-risks.jsp or from your SG representative. All opinions and estimates are given as of the date hereof and are subject to change. Société Générale is not obliged to inform investors of any change to such opinions or estimates.
This communication is exclusively directed to investors acting for their own account and categorised as eligible counterparties or professional clients as defined by 2004/39/EC Directive on markets in financial instruments.
This communication is confidential and use of this communication with or by any other party is prohibited. No part of it may be reproduced, distributed or transmitted without the prior written permission of SG.
Société Générale is a French credit institution (bank) that is supervised by the European Central Bank (‘ECB’) and the Autorité de Contrôle Prudentiel et de Résolution (the French Prudential Control and Resolution Authority) (‘ACPR’) .
NOTICE TO U.S. INVESTORS: This document is intended only for qualified institutional investors under SEC Rule 15(a) 6. Any U.S. person wishing to discuss this document or effect transactions in any security or financial instrument discussed herein should do so with or through SG Americas Securities LLC.
IRS CIRCULAR 230 DISCLOSURE: SG does not provide tax advice. Please note that (i) any discussion of US tax matters contained in this communication cannot be used by you for the purpose of avoiding tax penalties; (ii) this communication was written to support the promotion or marketing of the matters addressed herein; (iii) you should seek advice based on your particular circumstances from an independent tax advisor.
If you are a “U.S. person” (as defined by the U.S. Commodity Futures Trading Commission), please visit http://swapdisclosure.sgcib.com for important information with respect to derivative products. By transacting with Societe Generale, you are deemed to acknowledge that you have read and accepted the information provided on the website.
NOTICE TO RECIPIENTS IN HONG KONG: This document is issued solely to “professional investors” within the meaning of the Securities and Futures Ordinance (Cap. 571) of Hong Kong and any rules made under that Ordinance.
NOTICE TO RECIPIENTS IN SINGAPORE: This document is distributed by MAS licensed representatives of Societe Generale Singapore Branch and may only be provided to accredited investors, expert investors and institutional investors, as defined in Section 4A of the Securities and Futures Act. If you wish to discuss this document or effect transactions in any security discussed herein, you should do so with or through MAS licensed representatives of Societe Generale Singapore Branch. If you are an accredited investor or expert investor, please be informed that in our dealings with you, we are relying on certain exemptions to the Financial Advisers Act (“FAA”) - (1) the exemption in Regulation 33 of the Financial Advisers Regulations (“FAR”), which exempts us from complying with Section 25 of the FAA on disclosure of product information to clients; (2) the exemption set out in Regulation 34 of the FAR, which exempts us from complying with Section 27 of the FAA on recommendations; and (3) the exemption set out in Regulation 35 of the FAR, which exempts us from complying with Section 36 of the FAA on disclosure of certain interests in securities.
NOTICE TO RECIPIENTS IN JAPAN: This document is distributed in Japan by Societe Generale International Limited, which is authorised and regulated by the Financial Conduct Authority. This document is intended only for the Professional Investors as defined by the Financial Instruments and Exchange Act in Japan and only for those people to whom it is sent directly by Societe Generale International Limited, and under no circumstances should it be forwarded to any third party.
NOTICE TO RECIPIENTS IN EMEA: Societe Generale International Limited (“SGIL”) is a wholly owned subsidiary of Société Générale. SGIL is authorised and regulated by the Financial Conduct Authority. SGIL do not deal with, or for, Retail Clients (as defined by the 2004/39/EC Directive on markets in financial instruments). SGIL is a Swap Dealer registered with the CFTC and NFA.
Due to international regulations not all financial instruments/services may be available to all clients. You should be aware of and observe any such restrictions when considering a potential investment decision.
© 2017 Société Générale and its affiliates. All rights reserved.