Trend Following and Drawdowns: Is This Time Different?

The author would like to thank Harry Moore for his contribution to the analysis presented in this paper.

Is this time different?

This is the perennial question asked by investors experiencing a drawdown. And it is a question at the forefront of the minds of systematic traders. The haunting, nagging worry that some fundamental dynamic of markets has changed. That whatever the lessons of the past, this time it really is different: the inefficiency that your trading strategy once profited from has been wiped out, or negated due to some pervasive new law of markets.

And this worry isn’t without justification. There are several systematic strategies that were profitable years ago, but lost their edge as investors piled in. Take, for example, the classic equity pairs trading system. You identify pairs or baskets of stocks, find small perturbations away from the basket average and bet on those to revert. If I were to create a system that traded one-month mean-reversion, neutralised to all risk factors, you would have thought it a Ponzi scheme in the 1960s, when computers powerful enough to run the strategy didn’t really exist. It still worked remarkably well throughout the 1990s (when they did). But the edge disappeared through time – equity markets were becoming more efficient; the more capital that chased the trade, the less the inefficiency manifested itself. What was good news for large pension funds (who caused the inefficiencies by clumsy trading patterns) was bad news for systematic hedge funds seeking to make a profit. And I know, as I traded a pairs strategy in the early 2000s and saw the once strong performance fade. The options then were either to commit large resources to trading technology in a never-ending quest for speed, or to look for more accommodating pastures.

Figure 1. Account curve of systematic pairs trading (top) and fast trend following system (bottom) illustrate declining efficacy over time

Problems loading this infographic? - Please click here

Problems loading this infographic? - Please click here

Source: Man Group Database. Pairs strategy for illustration only, using equity sector mean reversion, rolling Z-score on normalised prices with a holding period less than 10 days. Returns before costs. Date range: June 1973 to May 2025. Fast trend strategy for illustration only. Trend system uses a moving average crossover with peak weight at two days, lagged by one day to simulate costs on a diversified set of approximately 100 markets. Date range: August 1960 to May 2025.

In the macro space, short-term trend following is another strategy where the lucrative returns of yesteryear have vanished into the efficiency of modern markets. Figure 1 (bottom chart) shows a trend system sensitive to trends shorter than one week plotted back to 1960. What explains this decline? The answer lies in the evolution of market dynamics. News gets priced in more quickly as technology advancements allow for more information to be digested instantaneously. We shouldn’t expect this anomaly to return – it is an artefact of an era in which information barriers, as well as liquidity barriers, existed and these have eroded over time.

Strategies that have stood the test of time

While some strategies have seen their edges eroded by advancing technology and market efficiency, others have demonstrated resilience over time. These strategies offer valuable lessons about the persistence of certain inefficiencies, even as markets evolve.

Traditional equity factors

Traditional quant equity trading strategies, where stocks are sorted based on fundamental characteristics, such as earnings ratios, have stood the test of time and the march of technology rather better. They have also illustrated that inefficiencies can manifest themselves in different ways. Investors as far back as Benjamin Graham in 1949 identified that undervalued stocks tend to outperform in the long run, as investors tend to overpay for growth potential. This inefficiency is significant and pervasive – and it can be used to pursue handsome returns in the long run - but it is not immune to significant drawdowns. This is partially attributable to investor behaviour: flows into the space can enhance portfolio returns, and outflows (which can be faster) have the opposite effect. Managers in this space have become more sophisticated over time, looking to diversify their sources of return and generating impressive return streams in the process. But drawdowns can and do happen – with the latest cross-industry ‘winter’ between 2018 and 2020 resulting in significant outflows, and a notable drag on returns. Patient investors here were able to weather that storm (and some opportunistic ones even size up their positions) to look to benefit from the inevitable rebound in performance as markets stabilised.

Time series trend following

Time series trend following is at the heart of the pain in the current environment of whipsaws, and is itself subject to that drawdown induced question – is this time different? Arguments for the prosecution sound compelling, with a world of firsts to point to: US government policy seeking to reverse the effects of globalisation (and US economic hegemony), widespread tariffs upending the free trade principal upon which economies and markets have relied, and an unpredictable US administration playing a game of chicken with markets and creating a very difficult environment in which trades can be sucked in one way and spat out the other. On top of this are questions about crowding and the behaviour of other market participants: has trend following become a fools’ errand?

In defence of trend following

The defence, therefore, is called upon to provide some evidence to back its case. And this strikes at the heart of making sensible investment decisions: gather as much information as you can, analyse, dissect, draw some inferences, and then step back and manage the odds in your favour.

We follow six lines of investigation, demonstrating that trends are persistent, the recent negative performance is to be expected at some point as part of a long track record and typically favourable outcomes have followed for those who have stayed the course.

1. How does the recent past compare to the distribution of strategy returns over time? Should we expect a drawdown of this magnitude?

To 30 April 2025, the rolling 12-month return for trend following is -18.6%. To put this into context, only twice has the Société Générale (SG) Trend index lost more than 15% over rolling 12-month windows, with the previous worst return at -16.0%, occurring in January 2019. These points are shown in the red highlighted bar in Figure 2.

What is notable about this distribution is how infrequent losses of greater than 15% are, while gains of more than 30% have been somewhat common. This highlights trend following’s inherent ability to risk manage itself and explains the positively skewed expected return profile. Interested readers can read more about this here.

Figure 2. Histogram of rolling 12-month SG trend returns. Red bar includes current drawdown

Problems loading this infographic? - Please click here

Source: SocGen and Man Group Database. Date range: January 2000 to April 2025.

That’s all well and good, but we are currently in a drawdown, and it’s painful. The next logical question to ask is what happened after the previous low in 12-month returns and were investors who stuck with the strategy rewarded?

The answer is a resounding yes, with returns of +82.5% from the January 2019 low to the start of this current 12-month drawdown in May 2024. However, that is just one data point. In the top chart in Figure 3, we consider all drawdowns greater than 10% over a rolling 12-month period and see what the return in the following 12 months has historically been. We see an average return of +9.8%.

Another avenue is to examine past periods of strong performance and the conditions that preceded them. The bottom chart in Figure 3 highlights rolling six-monthly windows of the best trend-following returns, showing that these periods were often preceded by softer patches. While I cannot claim to predict the future in 2025, history suggests that challenging phases for trend-following performance have frequently been followed by more favourable environments.

Figure 3. 12-month returns following all previous 12-month rolling drawdowns of more than 10% for trend following (top). Returns six months prior to trend following’s best six monthly returns (bottom)

Problems loading this infographic? - Please click here

Source: SocGen. Two rolling drawdown periods of more than 10% do not yet have a post 12-month return. These are highlighted in the left-hand chart. Date range: January 2000 to April 2025.

It is also worth considering how likely a drawdown of this magnitude is over a long-term track record. In Figure 4, we consider a hypothetical strategy with a respectable 0.5 Sharpe ratio targeting an annualised return volatility of 10% and zero correlation to traditional assets. This is a strategy many institutional investors find compelling. However, as shown, over a 25-year period (akin to today from the inception of the SG Trend index) the probability of hitting a 20% or greater drawdown is nearly four in five, so highly likely.

Figure 4. Estimated probability of reaching a two standard-deviation (20%) drawdown assuming a 0.5 Sharpe ratio, 10% annualised volatility and 0 autocorrelation over different time periods

Source: Man Group Database. Returns are assumed normal and identically distributed.

2. Is the strategy becoming overcrowded, or ‘arbitraged away’?

Estimating the crowdedness of a trade or trading strategy is a notoriously difficult endeavour, although there are clues. The most obvious place to look is assets under management and trading volumes, where you’d expect to see your percentage share rise through time as a space got more crowded. This is plotted below in Figure 5.

Figure 5. CTA AUM as a percentage of total hedge fund AUM (top), CTA participation in futures markets as a percentage of total futures volumes (bottom)

Problems loading this infographic? - Please click here

Problems loading this infographic? - Please click here

Source: Man Group database, BarclayHedge and fia.org. Date range: December 1997 to December 2024. Trend following and hedge fund AUM have been derived from the BarclayHedge database, Participation based on the total number of listed futures contracts traded by trend followers.

As shown, there is no evidence to suggest trend AUM is getting too big or becoming too large of a share of futures markets to raise cause for concern.

Another way to answer this question is to correlate performance with flow data. In theory, if a trade or strategy is crowded, then flows in or out of the space could be expected to have an outsized impact on performance, with flows in providing a tailwind and flows out a drag. As shown in Figure 6, we see no evidence of this occurring and find a correlation between flows and trend-following performance of -0.1. This relationship holds if using lagged performance, or absolute flows too, indicating it is robust.

Figure 6. Comparison of trend-following performance by month alongside normalised net flows per month. The correlation is -0.1, highlighting it is unlikely performance has been driven by flows

Problems loading this infographic? - Please click here

Source: Man Group Database, SocGen, Bloomberg. Date range: April 1973 to April 2025.

3. Is the environment no good?

There have been periods of sideways trend-following performance in the past where accusations abounded that the strategy was broken. These have historically proven unfounded.

The post-Global Financial Crisis (GFC) period was characterised by a central bank put which provided a headwind to performance, with rates pinned near zero and global economic policy converging. From 2009 to 2013, the SG Trend index returned -1.8%, making it a difficult hold, but investors who remained saw standout performance in 2014 of +19.7%, driven by steadily rising bond prices which momentum models captured. I remember hearing doubts during this time — about whether trend following still worked — but those who stayed the course were vindicated. It’s a reminder that patience is valuable in systematic strategies. Bringing us more up to date, as we saw in Figure 3, those who held through 2018 experienced positive returns following this difficult patch.

In 2025, on/off US policy dynamics have created a whipsaw effect, which has hurt trend-following strategies. Can this continue indefinitely? Our view is it seems unlikely. Firstly, as markets adapt to this new normal, the initial big reactions and reversals typically subside as they are priced in. Secondly, if the US becomes a long-term unreliable trading partner, other countries and markets may begin to rewire themselves and new supply chains could begin to form. This has the potential to set up new and divergent trends in global markets, which trend-following strategies are poised to capture.

Trend following has also historically wobbled before delivering some of its most notable returns, as my team noted in a recent paper, Why Patience Matters During Market Stress. During the GFC, for example, the SG Trend index suffered a peak to trough drawdown of over 17%, before repositioning and delivering crucial diversification (Figure 7).

Figure 7. Performance of the SG Trend Index and S&P 500 in the three months prior and during the GFC

Problems loading this infographic? - Please click here

Source: Man Group database, Bloomberg, SocGen. Date range: July 2007 to December 2009.

The final way to look at the environment for trend is by getting granular and digging down to individual markets. Individual market returns to trend quite often follow a pattern of ‘small losses, followed by large gains’ due to the positive long-term skew characteristics of running trend on a market. As a human looking at returns, this can be infuriating. There are periods of protracted losses, and this quickly leads to questions over any single market’s efficacy and whether it deserves its place in the portfolio.

Perhaps the best recent example of this is cocoa, which is shown in Figure 8. For around five years, trend following cocoa left systematic investors exasperated. It did nothing but underperform. You can imagine conversations in 2022, ‘well clearly something is up in that market which means it doesn’t trend. There must be supply issues or chocolate has the wrong demand elasticity’. Then, from around the second half of 2023, cocoa started to trend, and it made back all those losses and then some, ending 2024 as one of the strongest performing markets in a diversified trend strategy.

Figure 8. Trend-following returns of over 100 markets, with cocoa highlighted in purple. A period of protracted losses was followed by a recovery of similar magnitude. Patience is required to hold the underperformers through these times

Problems loading this infographic? - Please click here

Source: Man Group Database. Trend-following representative returns for over 100 markets. Date range: January 2018 to March 2024.

Trend following as a strategy can behave like this from time to time. The frustration of protracted losses can lead to doubts about portfolios or individual markets, but history suggests that those who persist have historically participated in subsequent returns when trends inevitably emerge. It’s not that the environment is "no good" — it’s just testing your patience.

4. Enduring the pain of drawdowns

We’ve looked at the probability of hitting certain drawdown levels. But theoretically, how much more pain could there be?

Figure 9 considers an attractive investment strategy which has no correlation to markets, a 0.6 Sharpe ratio and a volatility of 15%. If you look where you end up after 30 years of investment, you’ve made over eight times your capital. But how many investors would have stayed the course? Between 2004 and 2008, there was a peak to trough drawdown of 32%, and investors remained underwater for a whole nine years. During this period, it’s the exact same strategy, with returns picked from the exact same expected distribution, but luck was not on your side. Contrast that with the period from 2020 to 2025, when a high performance streak was hit.

Figure 9. Illustrative track record generated from a 0.6 Sharpe ratio, 15% annualised volatility investment strategy

Source: Man Group Database. Returns are assumed normal and identically distributed.

Drawdowns don’t just test the numbers — they test the nerves. It’s during these periods we all feel the sharp pain of losses more intensely than the slow satisfaction of gains. The temptation to abandon the strategy is strongest in these moments, but history shows that those who persist through the discomfort have often been rewarded.

This highlights a critical point: drawdowns are not evidence of a broken system — they are an inherent feature of probabilistic investing. Luck plays a role in short-term results, but over the long term, the statistics speak for themselves. A skilled manager, with a robust trend-following process, is like playing with a loaded deck. There may be unlucky hands along the way, but the odds are likely in your favour over time.

5. All very well but my investment committee can’t stomach the drawdown! What is the opportunity cost?

We’ve shown trend following has historically done well following a drawdown, but what is the opportunity cost related to selling in these periods?

We can go back to our 0.6 Sharpe ratio strategy from Figure 9 and consider the opportunity cost if we’d implemented a drawdown rule. Here we assume that at a 20% drawdown level, we deallocate, and only re-enter the strategy when the drawdown is made back. We are thus trying to avoid the situation where this time it is truly different and it has stopped working. In Figure 10, we imagine we allocate to three of these strategies which are uncorrelated and implement this rule. What happens?

Figure 10. Illustrative track record generated from three 0.6 Sharpe ratio, 15% annualised volatility investment strategies which are uncorrelated

A drawdown rule is implemented to move to cash whenever a 20% or greater drawdown is realised and only reallocate once that strategy is back to flat (Dynamic Portfolio). This capital is assumed to be allocated to cash in the drawdown period. This is compared to an equal allocation (Consistent Portfolio). Both start at one third allocation to each strategy. Allocations are shown in the top panel, returns in the bottom panel.

Problems loading this infographic? - Please click here

Source: Man Group Database. Returns are assumed normal and identically distributed, selected from a distribution assuming a Sharpe ratio of 0.6 and annualised volatility of 15%. Weights start at one third each and are not redistributed. When the dynamic portfolio breaches the drawdown rule, the allocation moves to cash, until the drawdown is recovered.

As shown in the top panel, each sub strategy exceeds its 20% drawdown threshold at least twice in this 30-year period, meaning there are parts of the track record with material allocations to cash. While further drawdowns are avoided, subsequent rebounds are missed too. The bottom panel shows the cost of this protection, as the Consistent Portfolio, the same one from Figure 9, significantly outperforms the Dynamic Portfolio which has the drawdown rule.

So, there is noticeable opportunity cost as a lot of upside can be missed over the long run. This is just one example but if we run the experiment 10,000 times and see where we end up, the Consistent Portfolio outperformed the Dynamic (drawdown) Portfolio by an average of 2.2% per annum.

It is also worth looking at historic data. Pivoting back to our quant equity strategy from the section on the ‘strategies that have stood the test of time’ at the beginning of this paper, we can investigate this factor portfolio.

The quant equity strategy, based on factors, had a great run until it suffered a very large drawdown period in 2020. Rumours abounded that factors were dead and these voices were not silenced for over two years. In Figure 11, we use an equally weighted portfolio of four market neutral factors, namely: value, momentum, quality and low beta. A leverage factor of two is applied, which gets this portfolio to an annual return volatility of around 12%, so comparable to the other portfolios we’ve looked at so far.

As shown below, the period from 1 January 2010 to 31 March 2020 saw a healthy +67% return following this strategy, while being uncorrelated to equity markets. A sharp peak to trough drawdown of -29% followed, wiped out a chunk of those gains and led to questions about the long-term viability of factor investing. However, those who held or even added saw their decisions vindicated, with a recovery of +84% from the lows to 30 April 2025.

Figure 11. Track record of an equity factor strategy which combines market neutral value, momentum, quality and low beta in equal proportions, leveraged two times

Source: Man Group Database, Bloomberg. Date range: 1 January 2010 to 30 April 2025. Indices combined are Dow Jones U.S. Thematic Market Neutral value, momentum, quality and anti beta.

Again, this highlights the importance of the underlying investment beliefs on the investment strategy and the manager’s skill to implement that strategy. Luck plays a role in short-term results, but over the long term, persistent market inefficiencies can play out.

6. Understanding the nature of trend following, and the role it plays

Finally, we should remind ourselves of why we hold trend following in portfolios in the first place. Of course, we know trend is diversifying to equity markets, but then so are other assets like bonds or multi-strategy hedge funds. The true utility of trend comes in the way that diversification is delivered, which is regularly referred to as ‘convexity’.

Convexity in this sense means trend following’s best performance has historically come in the worst periods for equity markets. In short, it’s diversification when you need it most (Figure 12).

Why it plays so nicely with equities, bonds or alternative strategies is down to this property. Trend is historically negatively correlated with markets when they’re selling off, not just uncorrelated. As shown, bonds or other alternative strategies do not share this property.

Figure 12. Quarterly performance of trend following (left), bonds (middle) and multi-strategy hedge funds (right) on y axis plotted against quarterly equity returns on x axis

Problems loading this infographic? - Please click here

Source: Man Group database, Bloomberg. Date range: January 2000 to March 2025 for trend following and bonds, January 2014 to March 2025 for multi-strategy. Trend following illustrated by SG Trend index, bonds by Barclays global aggregate and multi-strategy by HFRI Multi-Manager Pod Shop index.

What’s interesting about these plots is that bonds or other typical hedge fund approaches (we’ve illustrated multi-strategy using the HFRI Multi-Manager Pod Shop Index to try and capture a broad range) are mixed in the left side of the distribution. In short, when markets have really struggled, quite often these diversifiers don’t hold up. That’s where trend can really earn its bread. A year like this one, with sideways markets, has played more favourably to other diversifiers, and trend has struggled.

April 2025 was a clear example of the challenges trend following can face during sharp policy reversals. After the market shock caused by the crippling tariffs announced on 2 April, trend-following strategies repositioned and were ready to profit from further market declines as the fallout accelerated. On the morning of 9 April, a week later, markets continued to sell off, trend-following strategies had repositioned and were delivering positive returns. Headlines in the Financial Times live blog painted a grim picture: “Stock sell-off intensifies as China doubles down on retaliation”, “US fixed income volatility hits 18-month high”, “US corporate debt risk gauge climbs to highest level in two years” and “Gold prices surge as investors seek shelter from intensifying trade spat”. It was shaping up to be a dismal time for equity and bond investors.

Then, at lunchtime Eastern Time, a 90-day pause was announced which meant markets rebounded sharply. The next headline reads: “US stocks soar after Trump announces tariff pause on non-retaliating countries”. Equities had traced a classic “V”-shaped pattern.

But what if the tariff pause had not occurred? The counterfactual scenario is important to consider. By 8 April, the S&P 500 had already fallen 15% year-to-date, and intraday on 9 April, markets were continuing to decline. Without the tariff announcement, the sell-off would likely have deepened, leaving equity markets in crisis territory and trend following potentially set to gain. Instead, by June 2025, the S&P 500 is up year-to-date, with European and select Asian indices also showing gains — far from a market crisis.

Trend following thrives when markets experience sustained dislocations or crises. If recessionary pressures mount, something cracks in the Treasury market, or tariff barriers are reinstated at high levels, trend following is one of the few strategies in a portfolio that can be capable of delivering gains. Its adaptability to capture global trends — whether in bonds, currencies, commodities, or equities — has historically helped navigate periods of significant economic stress and protracted market drawdowns.

We’ve discussed the attractiveness of trend following as a long-term allocation here, with further articles here.

Conclusion

Drawdowns are painful. Even if you have complete confidence in the statistical properties of a system, and you know the right call for your portfolio is to hold on, drawdowns make that call harder. They can test patience, they make you feel returns more acutely, and they can lead you to question (and be questioned on) the rationale of an investment decision.

It is the job of our research teams to do all they can to improve the properties of our systems and hence limit the severity of drawdowns relative to expected returns. However, they are a necessary condition of most types of trading system, hence understanding their drivers and properties in relation to the expected distribution of returns can help separate the signal from the noise. None of us can predict the future, but we can utilise all the information available to us to make rational decisions.

All of which leads us to the following conclusions: no, we cannot guarantee trend following won’t be hurt by Trump’s posts on Truth Social, or any other kind of unforeseen event that prompts a reversal. But we can guarantee that we stand by the properties of a trend-following system. As a long-established trend follower who has endured many ups and downs, I’ve been asked many times over the years: is it different this time? The answer? It’s always different, the world is constantly changing, but as far as trend following is concerned – no, we don’t believe it’s different this time. We continue to learn, to evolve, and to try to improve every element of our trading systems. We also ultimately respect the nature of the markets, and admit that most of the time we just don’t know what’s going to happen. With that in mind, we remain convinced that the best approach is to stick to our process: trend following has produced return streams that have historically benefitted investors who can stay the course.

With thanks to Harry Moore, Tom Bowles, Yash Panjabi, Max Buchanan and Claudia Cilleruelo Pascual for their contributions and assistance in preparing figures.

Bibliography

Robertson, G, Goodall, R. (2023), “Trend-Following: A Different Point of Skew”, Man Institute, Available at: https://www.man.com/maninstitute/trend-following-different-point-skew

Van Hemert, O., Ganz, M., Harvey, C. R., Rattray, S., Sanchez Martin, E. and Yawich, D., “Drawdowns”, Journal of Potfolio Management, Available at: https://eprints.pm-research.com//17511/37716/index.html?40813

Abou Zeid, T., Panjabi, Y. and Moore, H. (2025), “Why Patience Matters During Market Stress”, Man Institute, Available at: https://www.man.com/insights/why-patience-matters-during-market-stress

Robertson, G. (2023), “What's Trending: Trend-following - What's Not to Like?”, Man Institute, Available at: https://www.man.com/maninstitute/trend-following-what-not-to-like

Korgaonkar, R. (2025), “The Big Picture: Divergent Policies, Divergent Markets”, Man Institute, Available at: https://www.man.com/insights/the-big-picture-divergent