Please Share:: 
India Equity Research Reports, IPO and Stock News
Visit http://indiaer.blogspot.com/ for complete details �� ��
Investment philosophy: Passive versus active management
One of the most common debates in modern investment philosophy revolves around whether
active management of equities is efficient or investors should be advised to passively
manage their portfolios. Investors have realized that traditional investment vehicles become
increasingly risky when the markets take a rollercoaster ride. The traditional long only
investment styles gets handcuffed in volatile market as neither it safeguards the investments
nor is able to take the advantage of the short side of the market. Addressing this issue, we
introduce quantitative concept of co-integration based optimization that deals with issues
such as:
A classic index tracking strategy
A long-short equity market neutral strategy
By using the co-integration based classic index tracking strategy, portfolio managers (PMs)
can replicate a benchmark in terms of returns and volatility. With the long-short equity
market neutral strategy, PMs can seek to minimize volatility and generate steady returns
under all market circumstances. Combination of both could enable PMs to enhance the
properties of basic strategies.
Various portfolio enhancement strategies
Investors and PMs always quest to achieve optimal returns by enhancing the portfolio
through various strategies. The spectrum of such strategies is boundless. Each portfolio has
its unique characteristics and, hence, the strategy has to be unique. Few of the enhancement
strategies are:
Co-integration optimization approach
Statistical long/short market neutral strategy
Enhance indexing
Leveraging and structured instruments
Superior stock selection
The idea of using enhancement strategies is either to outperform the benchmark what we
call Alpha over market or use these strategies to derive consistent returns with minimum
volatility i.e. the absolute return strategy. Enhanced indexing aims at gaining alpha over the
benchmark whereas long-short market neutral strategies are used for absolute returns. Cointegrating
optimization helps us achieve the best of both worlds.
Introduction to co-integration
Co-movement in market index and stock prices are seen as the effect of common underlying
economic factors such as macroeconomic conditions (both domestic fundamentals and
international economic developments) and investors’ expectations and behaviour. The
primary objective of co-integration is to efficiently estimate the degree of co-movements in
stocks and market index using available market information. Co-integration uses historical
price patterns and long-run equilibrium relationship to project the future performance. Cointegration
checks long-term relationship between the market index and portfolio stocks, and
exploits it to optimize the portfolio weights to enhance portfolio returns.
Co-integration on the Nifty: Using co-integration, PMs can select the basket of stocks
from Nifty components that are highly co-integrated with the Nifty index, and apply the cointegration
techniques to determine the optimized portfolio weights to enhance portfolio
returns.
Co-integration techniques are applied to replicate benchmark index with minimum tracking
error. It can be further applied in long-short market neutral strategy to generate steady
returns with minimum volatility.
Types of co-integration optimization based strategies
A classic index tracking strategy: This strategy aims to replicate the benchmark index.
Tracking portfolios are expected to have similar returns, volatility and high correlation with
the index. Index tracking strategy is a structured approach to index funds investing, which
lies somewhere in between active and passive management styles (though it is closer to
indexing than to active management). Similar to the passive indexing strategy, index
tracking gives special emphasis to portfolio diversification.
A long-short equity market neutral strategy: This strategy aims to generate returns at
the desired alpha over benchmark with less correlation with the benchmark returns. As seen
historically, lesser Alpha over benchmark strategy generates more consistent returns with
low volatility and less correlation. Whereas, higher the desired alpha over benchmark cointegration
relationship could begin to break down and consequently market neutral strategy
can become more volatile. Using the above strategy principle, we have back-tested
Edelweiss portfolio optimizer (EPO) long-short strategy, where we have tried to achieve
absolute consistent returns. Please refer page 5 for details on strategy back-test returns.
Back testing
Back testing methodology for long-short equity portfolio
Create two synthetic indices taking the 12-month trailing calibration period
NIFTY_Plus to generate returns of NIFTY + a (i.e. NIFTY + 50%)
NIFTY_Minus to generate returns of NIFTY - a (i.e. NIFTY - 50%)
So Portfolio_Plus would aim to track NIFTY + 50% return series whereas
Portfolio_Minus will track NIFTY - 50% returns series
Generate portfolio which consists of top 15 stocks by weight in NIFTY on last
trading day of the month
Determine weights of the stocks in the portfolio by regressing Log (NIFTY_Plus)
versus log of portfolio stocks over trailing one-year and repeated the same for
NIFTY_Minus tracking portfolio stocks
Y(+ve) = L1X1 + L2X2 + ------------- L15X15 + C1
Y(-ve) = S1X1 + S2X2 + ------------- S15X15 + C2
Where- Ln = Co-efficient of nth stock in long portfolio
Sn = Co-efficient of nth stock in short portfolio
Cn = Constant
Ensure the portfolio created is co-integrated with the synthetic NIFTY by testing for
stationarity in the residuals of above regression.
The co-integration coefficients L1, L2 …. , Ln are normalized to sum to one to give
optimized portfolio weights. A negative weight indicates taking a short position on
the portfolio stock.
Difference of Portfolio_Plus and Portfolio_Minus weights would give the final longshort
portfolio weights for the current month
Y(L-S) = ln(L1-S1)X1 + ln(L2-S2)X2 + ------------- ln(L15-S15)X15 + C3
The model has been back-tested on the historical data for the last eight years
At the end of the last trading day of the month, portfolio weights are tweaked by
regressing synthetic NIFTY ( NIFTY_Plus and NIFTY_Minus) on the portfolio stocks
over trailing one year and tested for co-integration in the new portfolio of stocks
India Equity Research Reports, IPO and Stock News
Visit http://indiaer.blogspot.com/ for complete details �� ��
Investment philosophy: Passive versus active management
One of the most common debates in modern investment philosophy revolves around whether
active management of equities is efficient or investors should be advised to passively
manage their portfolios. Investors have realized that traditional investment vehicles become
increasingly risky when the markets take a rollercoaster ride. The traditional long only
investment styles gets handcuffed in volatile market as neither it safeguards the investments
nor is able to take the advantage of the short side of the market. Addressing this issue, we
introduce quantitative concept of co-integration based optimization that deals with issues
such as:
A classic index tracking strategy
A long-short equity market neutral strategy
By using the co-integration based classic index tracking strategy, portfolio managers (PMs)
can replicate a benchmark in terms of returns and volatility. With the long-short equity
market neutral strategy, PMs can seek to minimize volatility and generate steady returns
under all market circumstances. Combination of both could enable PMs to enhance the
properties of basic strategies.
Various portfolio enhancement strategies
Investors and PMs always quest to achieve optimal returns by enhancing the portfolio
through various strategies. The spectrum of such strategies is boundless. Each portfolio has
its unique characteristics and, hence, the strategy has to be unique. Few of the enhancement
strategies are:
Co-integration optimization approach
Statistical long/short market neutral strategy
Enhance indexing
Leveraging and structured instruments
Superior stock selection
The idea of using enhancement strategies is either to outperform the benchmark what we
call Alpha over market or use these strategies to derive consistent returns with minimum
volatility i.e. the absolute return strategy. Enhanced indexing aims at gaining alpha over the
benchmark whereas long-short market neutral strategies are used for absolute returns. Cointegrating
optimization helps us achieve the best of both worlds.
Introduction to co-integration
Co-movement in market index and stock prices are seen as the effect of common underlying
economic factors such as macroeconomic conditions (both domestic fundamentals and
international economic developments) and investors’ expectations and behaviour. The
primary objective of co-integration is to efficiently estimate the degree of co-movements in
stocks and market index using available market information. Co-integration uses historical
price patterns and long-run equilibrium relationship to project the future performance. Cointegration
checks long-term relationship between the market index and portfolio stocks, and
exploits it to optimize the portfolio weights to enhance portfolio returns.
Co-integration on the Nifty: Using co-integration, PMs can select the basket of stocks
from Nifty components that are highly co-integrated with the Nifty index, and apply the cointegration
techniques to determine the optimized portfolio weights to enhance portfolio
returns.
Co-integration techniques are applied to replicate benchmark index with minimum tracking
error. It can be further applied in long-short market neutral strategy to generate steady
returns with minimum volatility.
Types of co-integration optimization based strategies
A classic index tracking strategy: This strategy aims to replicate the benchmark index.
Tracking portfolios are expected to have similar returns, volatility and high correlation with
the index. Index tracking strategy is a structured approach to index funds investing, which
lies somewhere in between active and passive management styles (though it is closer to
indexing than to active management). Similar to the passive indexing strategy, index
tracking gives special emphasis to portfolio diversification.
A long-short equity market neutral strategy: This strategy aims to generate returns at
the desired alpha over benchmark with less correlation with the benchmark returns. As seen
historically, lesser Alpha over benchmark strategy generates more consistent returns with
low volatility and less correlation. Whereas, higher the desired alpha over benchmark cointegration
relationship could begin to break down and consequently market neutral strategy
can become more volatile. Using the above strategy principle, we have back-tested
Edelweiss portfolio optimizer (EPO) long-short strategy, where we have tried to achieve
absolute consistent returns. Please refer page 5 for details on strategy back-test returns.
Back testing
Back testing methodology for long-short equity portfolio
Create two synthetic indices taking the 12-month trailing calibration period
NIFTY_Plus to generate returns of NIFTY + a (i.e. NIFTY + 50%)
NIFTY_Minus to generate returns of NIFTY - a (i.e. NIFTY - 50%)
So Portfolio_Plus would aim to track NIFTY + 50% return series whereas
Portfolio_Minus will track NIFTY - 50% returns series
Generate portfolio which consists of top 15 stocks by weight in NIFTY on last
trading day of the month
Determine weights of the stocks in the portfolio by regressing Log (NIFTY_Plus)
versus log of portfolio stocks over trailing one-year and repeated the same for
NIFTY_Minus tracking portfolio stocks
Y(+ve) = L1X1 + L2X2 + ------------- L15X15 + C1
Y(-ve) = S1X1 + S2X2 + ------------- S15X15 + C2
Where- Ln = Co-efficient of nth stock in long portfolio
Sn = Co-efficient of nth stock in short portfolio
Cn = Constant
Ensure the portfolio created is co-integrated with the synthetic NIFTY by testing for
stationarity in the residuals of above regression.
The co-integration coefficients L1, L2 …. , Ln are normalized to sum to one to give
optimized portfolio weights. A negative weight indicates taking a short position on
the portfolio stock.
Difference of Portfolio_Plus and Portfolio_Minus weights would give the final longshort
portfolio weights for the current month
Y(L-S) = ln(L1-S1)X1 + ln(L2-S2)X2 + ------------- ln(L15-S15)X15 + C3
The model has been back-tested on the historical data for the last eight years
At the end of the last trading day of the month, portfolio weights are tweaked by
regressing synthetic NIFTY ( NIFTY_Plus and NIFTY_Minus) on the portfolio stocks
over trailing one year and tested for co-integration in the new portfolio of stocks
No comments:
Post a Comment