ݺߣ

ݺߣShare a Scribd company logo

Rho research trend-following is dead long live trend-following

•
•
John Robertson
John Robertson

Trend-following strategies have been unprofitable in recent years due to a lack of strong trends in most markets. Prices have remained trapped within narrow ranges for prolonged periods, making it difficult for trend-following to be successful. The last 3 years have been the most challenging for trend-following since 1995, with fewer than 45% of markets exhibiting trends. However, trend-following is not dead - strong trends often follow long periods of consolidation, and will likely emerge again when market conditions change to allow for smooth price moves between levels.

1 of 7
Downloaded 22 times
  Trend‐following is Dead! Long live trend‐following!    Is trend‐following dead?  The last three years have not been profitable for trend‐followers.  Does this mean trend‐following is  dead?  The short answer is no.  For a longer answer, we quote Stanley Fink, CEO of International  Standard Asset Management, ex‐CEO of Man Group, PLC (via Bloomberg.com, Jesse Westbrook  reporting, Oct. 3, 2013): “People say to me ‘Isn’t trend‐following dead?’” Fink said. “I probably  remember four or five occasions in my experience in the fund‐management world when people have  said that. In each case, it’s been absolutely wrong, and the industry tends to roar back when most  people have capitulated.”     Why has trend‐following not worked? (Because prices have remained in a narrow range for long time)  Trend‐following works when a large percentage of markets in a portfolio have smooth price moves from  one level to another.   This means trend‐following is unprofitable if there are few markets with smooth  trends, if price moves are short, or if price moves are choppy, i.e., they go back and forth rapidly.    Fundamentally, if your average trade length, or holding period, is x‐days, each trade is a forecast that  prices will move substantially in the direction of the trade at the end of x‐days.  Trend‐following buys  strength and sells weakness. This means trend‐following systems buy markets that are rising and sell  markets that are falling.  Each trade is a forecast that prices will be higher (or lower) at the end of the  holding period.  The trade will be unprofitable if prices reverse course over the x‐day period, falling  when higher prices are forecast, or rising when lower prices are predicted.     When do prices reverse instead of following through in the desired direction?  When there is confusion  and volatility, as has been the case after the Lehman bankruptcy in September, 2008.      Is there a way to visualize prices trapped in a narrow range? (i.e., How to identify consolidations?)  A price chart is a simple way to visualize price action.  When prices are trapped in a triangle or in a  rectangle for long periods of time, then they reverse at the boundaries of that formation.  However, by  design, trend‐following systems are forecasting that they will break‐out of the consolidation pattern.   Hence, trend‐following systems will be unprofitable when prices are trapped in a consolidation pattern.  We illustrate these ideas using the Cocoa futures prices chart in Figure 1 below.  Observe how prices  were inside a large triangle lasting nearly three years.  Within the large triangle, there are many periods,  shown using dotted rectangles, when prices traded in a narrow range, sometimes within a smaller  Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch 
  triangular pattern. Trend‐following strategies that traded the breakout were unprofitable until prices  staged a successful breakout.  Short move after breakout, into larger resistance.  Successful Break‐out  Sustainable breakout?  False Break‐out     Figure 1: This continuous chart of Cocoa futures contract shows a series of triangular consolidations and breakouts from those consolidations. Observe how prices have stayed within the large triangle for almost three years before breaking out in July/August this year. Trend‐following strategies that bought near the upper triangular line, or sold near the lower triangular line were unprofitable due to reversals. The dashed circles show some, but not all, of the points at which break‐out systems could have generated unprofitable signals.   Trend‐following models come in many flavors and styles, varying in speed with which they react, so that  the holding period, or length of trade, can be varied. Figure 1 also shows some, but not all, of the points  at which trend‐following strategies could have generated unprofitable signals.  Figure 1 shows the  essential problem trend‐following strategies have faced over the last few years.  A majority of markets  have consolidated for a very long time.  This means when they trend, opportunities will be plentiful.    How bad is it? (Really bad)  We look at the Rho Trend Barometer over rolling 36‐month intervals starting 1995 to show that trading  conditions have never been this bad, at least since 1995.  The Rho Trend Barometer measures the  percentage of markets with measurable trends on the last trading day of each month.  The break‐even  Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch 
  level for trend‐following profitability is approximately 45%, which means if fewer than 45% of market  are trending, then trend‐followers as a group are likely to be unprofitable.  70% > 60% ==>Period of strong, sustained trends for many months % of 36‐month intervals with %T > 45% 60% 50% 40% < 50% ==> Prolonged period with few trends: trend‐following  strategies are unsuccessful 30% 20% Extremely difficult for trend‐following strategies: only 33% of rolling 36‐ month intervals had meanigful trends 10% 1/1/1998 8/1/1998 3/1/1999 10/1/1999 5/1/2000 12/1/2000 7/1/2001 2/1/2002 9/1/2002 4/1/2003 11/1/2003 6/1/2004 1/1/2005 8/1/2005 3/1/2006 10/1/2006 5/1/2007 12/1/2007 7/1/2008 2/1/2009 9/1/2009 4/1/2010 11/1/2010 6/1/2011 1/1/2012 8/1/2012 3/1/2013 0% Figure 2: We show the long‐term evolution of the number of months with sustainable trends over rolling 36‐month intervals starting in 1995 (so first interval ends Dec, 1997). Between 1995 and 2009, there were very few 36‐month intervals in which fewer than half the months were below break‐even. Since 2011, this number has declined sharply, to about 33%, i.e., only 12 of 36 months had trends above break‐level of 45 percent. Trend‐followers have not faced such tough trading conditions since 1995.    On the vertical axis of Figure 2 we show the percentage of months (over rolling 36‐month intervals) with  trends above the 45% break‐even level.  Naturally, the greater the number of months with above‐ average trends, the more profitable trend‐followers will be.  Note how the number of months with  sustained trends has dropped to an all‐time low of only 33% during 2013.         Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch   
  Can there be great trends after consolidations?  We illustrate this idea using the Simex Nikkei futures contract.  This market consolidated in a triangular  formation for nearly four years, before staging a powerful breakout.  Thus, external conditions must  change before markets will have major trends.  In the case for the Nikkei, it was the election of Prime  Minister Abe and his inflationary policies that fundamentally changed the situation in Japan, and the  markets adjusted to the new reality with a major trend.      Figure 3: We show the long‐term chart of the Simex Nikkei contract to show that a change in the external factors (in this case a new prime minister) is needed to end the market consolidation, and that a break‐out after a prolonged consolidation can lead to powerful trends.    Trends have been limited to Bonds, and Stocks.  We show the Rho Trend Barometer by sector to show that over the last few years, trends have been  limited to Bonds (interest rate complex), and equity index futures.  We will also show that there have  been very few trends in commodities.  Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch 
  Financials Sector trend strength (<30 or > 30 is desirable) 90 60 30 0 Financials ‐30 ‐60 9/1/2013 5/1/2013 1/1/2013 9/1/2012 5/1/2012 1/1/2012 9/1/2011 5/1/2011 1/1/2011 9/1/2010 5/1/2010 1/1/2010 9/1/2009 5/1/2009 1/1/2009 9/1/2008 5/1/2008 1/1/2008 9/1/2007 5/1/2007 1/1/2007 ‐90 Figure 4: We show Rho Trend Barometer (RTB) for the Financials sector (interest rates and long bonds). There are good trends when the RTB is greater than 30 or less than ‐30. Note that most the RTB values have been greater than +30, implying good long‐side trends. Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch   
  Stock Indexes 90 30 0  Stock Indexes ‐30 ‐60 9/1/2013 5/1/2013 1/1/2013 9/1/2012 5/1/2012 1/1/2012 9/1/2011 5/1/2011 1/1/2011 9/1/2010 5/1/2010 1/1/2010 9/1/2009 5/1/2009 1/1/2009 9/1/2008 5/1/2008 1/1/2008 9/1/2007 5/1/2007 ‐90 1/1/2007 Sector trend strength (<30 or > 30 is desirable) 60   Figure 5: We show Rho Trend Barometer (RTB) for the Stock Index sector. There are good trends when the RTB is greater than 30 or less than ‐30. Note for example in 2008 when values were < ‐30 after the vicious sell offs following the Lehman bankruptcy. Also, observe the recent values > +30 due to strong moves in Simex Nikkei and Swiss Market Index. Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch 
  All Commodities 60 30 0 All Commodities ‐30 ‐60 9/1/2013 4/1/2013 11/1/2012 6/1/2012 1/1/2012 8/1/2011 3/1/2011 10/1/2010 5/1/2010 12/1/2009 7/1/2009 2/1/2009 9/1/2008 4/1/2008 11/1/2007 6/1/2007 ‐90 1/1/2007 Sector trend strength (<30 or > 30 is desirable) 90 Figure 6: We show Rho Trend Barometer (RTB) for the Commodities sector. There are good trends when the RTB is greater than 30 or less than ‐30. Note for example in 2008 when values were < ‐30 after the vicious sell offs following the Lehman bankruptcy. However, there have been few trends since 2011, with most of the values between ‐30 and +30, the neutral zone.      68% of markets have consolidated since 2010: So there is a lot of tinder for trends to catch fire  A recent internal review found that only 14/44 or approximately 32% of markets in our portfolio were  not locked in a long‐term consolidation pattern.  This means fully 68% of the markets have not trended  since 2010.  This explains why trend‐followers have not done well over the past few years.  However,  this also shows the potential for powerful breakouts as soon as external conditions improve  Past performance is not necessarily indicative of future results. There is significant risk of loss in trading derivative products.    © 2013 Rho Asset Management AG    Baarerstrasse 78    CH‐6300 Zug    Switzerland  www.rhoam.ch 

More Related Content

Rho research trend-following is dead long live trend-following