- Joined
- 11/25/12
- Messages
- 3
- Points
- 11
"Trend Following is one of the stock market's biggest con, I would even go so far to compare the concept to a cult like Scientology... I get the same sort of value from Trend Following as I do from supernatural operators such as Uri Geller and horoscope readers."
Famed quant trader Victor Niederhoffer is not a big fan of Trend Following...
His opinions (quoted above) illustrate the typical divide between two categories of traders: quants and technical analysts. Quants usually think of technicians as tea-leaf or goat-entrail readers - trying to predict the market in a witchcraft, archaic manner.
Let's put aside the fact that Niederhoffer blew up twice 1 to concentrate on his claim of lack of seriousness in Trend Following and its inherent invalidity.
There will always be trends!
Trend Following, in its general public perception, appears clearly anchored in the realms of Technical Analysis. This comes with a heavy baggage of "folk culture", and many fool's quests for that magical holy grail trend indicator or system. Not the greatest publicity for credibility.
One argument that most proponents of Trend Following like to put forward can be summarized as follows:
"There will always be trends. Without trends, there is no market". This is true. However, when concluding that: "therefore, Trend Following will always be profitable", the argument discredits itself by the lack of rigor in its logic.
It is not because markets exist, and therefore trend, that Trend Following is a profitable strategy. It is a necessary but insufficient condition. Random markets illustrate that point: traders interested in the quantitative and statistical analysis of the market know that random processes will exhibit a stronger trending behavior than intuitively anticipated - purely from the effects of chance.
Below is an output from a random price generator:
This simulated price chart presents a "textbook" moving average cross-over. Plenty of such examples can be found from a random generator. Yet, by definition, no trading system or strategy - including Trend Following - would make money off random markets.
This lack of rigor and over-simplifications could explain a dismissal of Trend Following by a large part of the quant community.
However, there is more to it than the "caricature" so often depicted, and Trend Following can still show potential to quant traders. The notable entry of very large quant fund AQR in the Managed Futures (2) space with the recent launch of their new Trend Following funds proves the point.
Trend Following basics
Trend Following is a strategy positioned to benefit from volatility and large "fat-tail" moves that exist in market distributions. Basic strategies are fairly simple and unsophisticated: they identify the start of trends using indicators such as moving average cross-overs, or channel breakouts, use a volatility-based stop to cut losses, and stay with the trends until they clearly reverse according to these same indicators. The expectancy of such strategy is characterized by a low win ratio counterbalanced with a high payoff, making it profitable on balance.
One of the keys in a successful Trend Following system lies in additional system functions such as position sizing and risk management. These are typically problems of a quantitative nature, more tractable mathematically speaking.
Economic and behavioral explanations for Trend Following
Behavioral finance has documented some of the investor biases explaining the rationale for Trend Following:
Time series analysis
Econometrics and time series analysis are clearly the quant's kingdom. Although some measures such as volatility, kurtosis, serial correlation clearly influence Trend Following, very little research has been published on the relationship between the performance of Trend Following strategies and price distribution characteristics or statistical and quantitative analysis of market time series.
There is a wide scope to research and build statistical filters based on the characteristics above, to identify more promising Trend Following trades and enhance the performance of a simple strategy.
Finally, the time series characteristics could be used to define different market regimes, with an adaptive Trend Following strategy, which would match the shifting regimes. The concept of trading regimes is a concept that has recently become popular in the quant blogsphere.
Dynamic Asset Allocation
An alternative way to look at Trend Following is to consider it as dynamic asset allocation method.
Mebane Faber's paper: A Quantitative Approach to Tactical Asset Allocation [2007] presents a simple quantitative method that improves the risk-adjusted returns across various asset classes. Effectively, Faber borrows a large part of his method to Trend Following, by considering a simple moving average timing model to move in and out of each asset class. This dynamic allocation model shows interesting empirical results: equity-like returns with bond-like volatility and drawdown.
Instead of a strategy on its own right, Trend Following can be used in the context of portfolio management to dynamically allocate capital to the various assets in the portfolio, which should reduce its risk and volatility.
A new exotic beta for portfolio optimization
Above is a diagram by Andrew Lo, illustrating how alpha morphs into beta over time, as an initial strategy becomes more common/popular.
Large CTAs involved in Trend Following tend to have strongly correlated results, which would support the idea that Managed Futures/Trend Following have become an "exotic beta". This is a concept discussed in Conquest Group's paper: The Beta of Managed Futures.
In the paper, the authors describe a simple mechanical method representative of Trend Following strategy (namely a multiple timeframe Donchian breakout system) and argue that it constitutes a relevant index for Trend Following. Studies of correlation between this index and large Trend Following funds show strong interdependence (correlation figures ranging from 0.75 to 0.9).
This implies that Trend Following returns can be easily replicated. Managed Futures could therefore be considered an "exotic beta" asset class to diversify a portfolio.
Conclusion
There is definitely more to Trend Following than goat-entrail reading. Trend Following can provide quants with research and trading material. We have seen that there are several ways to implement and use Trend Following, all of which are quantitative, whether it as a standalone trading strategy, as part of a market regime approach, a dynamic asset allocation method or even a new "exotic beta" asset class for portfolio management.
About the Author:
Jez Liberty is a System Trader and Developer with a keen interest on researching Trend Following from a quantitative angle He shares his research on the Au.Tra.Sy blog (Automated Trading System)
He has been living in London for the last eight years, working as an IT professional for software companies and the banking industry.
_______________________________________________________________________
References:
1: Niederhoffer suffered large losses in the Asian crisis of 1997 and had to close his fund: Niederhoffer Investments. Ten years later, during the credit crunch in September 2007, Matador Fund was closed after a decline in value of more than seventy-five percent<br />
2: the terms CTA (Commodity Trading Advisors), Managed Futures and Trend Following are usually used interchangeably.
Famed quant trader Victor Niederhoffer is not a big fan of Trend Following...
His opinions (quoted above) illustrate the typical divide between two categories of traders: quants and technical analysts. Quants usually think of technicians as tea-leaf or goat-entrail readers - trying to predict the market in a witchcraft, archaic manner.
Let's put aside the fact that Niederhoffer blew up twice 1 to concentrate on his claim of lack of seriousness in Trend Following and its inherent invalidity.
There will always be trends!
Trend Following, in its general public perception, appears clearly anchored in the realms of Technical Analysis. This comes with a heavy baggage of "folk culture", and many fool's quests for that magical holy grail trend indicator or system. Not the greatest publicity for credibility.
One argument that most proponents of Trend Following like to put forward can be summarized as follows:
"There will always be trends. Without trends, there is no market". This is true. However, when concluding that: "therefore, Trend Following will always be profitable", the argument discredits itself by the lack of rigor in its logic.
It is not because markets exist, and therefore trend, that Trend Following is a profitable strategy. It is a necessary but insufficient condition. Random markets illustrate that point: traders interested in the quantitative and statistical analysis of the market know that random processes will exhibit a stronger trending behavior than intuitively anticipated - purely from the effects of chance.
Below is an output from a random price generator:
This lack of rigor and over-simplifications could explain a dismissal of Trend Following by a large part of the quant community.
However, there is more to it than the "caricature" so often depicted, and Trend Following can still show potential to quant traders. The notable entry of very large quant fund AQR in the Managed Futures (2) space with the recent launch of their new Trend Following funds proves the point.
Trend Following basics
Trend Following is a strategy positioned to benefit from volatility and large "fat-tail" moves that exist in market distributions. Basic strategies are fairly simple and unsophisticated: they identify the start of trends using indicators such as moving average cross-overs, or channel breakouts, use a volatility-based stop to cut losses, and stay with the trends until they clearly reverse according to these same indicators. The expectancy of such strategy is characterized by a low win ratio counterbalanced with a high payoff, making it profitable on balance.
One of the keys in a successful Trend Following system lies in additional system functions such as position sizing and risk management. These are typically problems of a quantitative nature, more tractable mathematically speaking.
Economic and behavioral explanations for Trend Following
Behavioral finance has documented some of the investor biases explaining the rationale for Trend Following:
- Anchoring bias: tendency to rely too heavily on recent price history to estimate "fair-value".
- Herding and feedback trading: tendency for traders to act as a group and jump on the bandwagon of a rising prices trend.
- Confirmation bias: people tend to look for information supporting their beliefs and consider recent price moves to be representative of future prices. This leads investors to over-allocate to markets having already risen and under-allocate to fallen markets. This behaviour favors trend continuation.
- Overreaction: market participants overreact to new information, creating larger-than-warranted effect on market price and stronger trends.
Time series analysis
Econometrics and time series analysis are clearly the quant's kingdom. Although some measures such as volatility, kurtosis, serial correlation clearly influence Trend Following, very little research has been published on the relationship between the performance of Trend Following strategies and price distribution characteristics or statistical and quantitative analysis of market time series.
There is a wide scope to research and build statistical filters based on the characteristics above, to identify more promising Trend Following trades and enhance the performance of a simple strategy.
Finally, the time series characteristics could be used to define different market regimes, with an adaptive Trend Following strategy, which would match the shifting regimes. The concept of trading regimes is a concept that has recently become popular in the quant blogsphere.
Dynamic Asset Allocation
An alternative way to look at Trend Following is to consider it as dynamic asset allocation method.
Mebane Faber's paper: A Quantitative Approach to Tactical Asset Allocation [2007] presents a simple quantitative method that improves the risk-adjusted returns across various asset classes. Effectively, Faber borrows a large part of his method to Trend Following, by considering a simple moving average timing model to move in and out of each asset class. This dynamic allocation model shows interesting empirical results: equity-like returns with bond-like volatility and drawdown.
Instead of a strategy on its own right, Trend Following can be used in the context of portfolio management to dynamically allocate capital to the various assets in the portfolio, which should reduce its risk and volatility.
A new exotic beta for portfolio optimization
Large CTAs involved in Trend Following tend to have strongly correlated results, which would support the idea that Managed Futures/Trend Following have become an "exotic beta". This is a concept discussed in Conquest Group's paper: The Beta of Managed Futures.
In the paper, the authors describe a simple mechanical method representative of Trend Following strategy (namely a multiple timeframe Donchian breakout system) and argue that it constitutes a relevant index for Trend Following. Studies of correlation between this index and large Trend Following funds show strong interdependence (correlation figures ranging from 0.75 to 0.9).
This implies that Trend Following returns can be easily replicated. Managed Futures could therefore be considered an "exotic beta" asset class to diversify a portfolio.
Conclusion
There is definitely more to Trend Following than goat-entrail reading. Trend Following can provide quants with research and trading material. We have seen that there are several ways to implement and use Trend Following, all of which are quantitative, whether it as a standalone trading strategy, as part of a market regime approach, a dynamic asset allocation method or even a new "exotic beta" asset class for portfolio management.
About the Author:
Jez Liberty is a System Trader and Developer with a keen interest on researching Trend Following from a quantitative angle He shares his research on the Au.Tra.Sy blog (Automated Trading System)
He has been living in London for the last eight years, working as an IT professional for software companies and the banking industry.
_______________________________________________________________________
References:
1: Niederhoffer suffered large losses in the Asian crisis of 1997 and had to close his fund: Niederhoffer Investments. Ten years later, during the credit crunch in September 2007, Matador Fund was closed after a decline in value of more than seventy-five percent<br />
2: the terms CTA (Commodity Trading Advisors), Managed Futures and Trend Following are usually used interchangeably.
