Blank_Alttext
Blank_Alttext
Share |
   
Blank_Alttext
Blank_Alttext
blank
 
 
Home » ETF Model Portfolios » Research Paper: The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation

The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation!

This research article was awarded as Editor's Pick on Seeking Alpha! September 24th, 2013

 

Portfolio construction techniques based on predicted risk, without expected returns, have become quite popular within the last couple of years. Especially on Seeking Alpha, great articles have been published by fellow contributors, covering the full range of modern portfolio theory and other different kind of tactical asset allocation concepts.

 

However, most of those published articles are based on back testing, whereas limited historical data as well as highly unlikely similar future performances of certain asset classes (e.g. the 30 year bull market in bonds or the very nice trend structure of equities between 2000 and 2010) are making an evaluation or comparisons of different kind of portfolio construction techniques quite difficult. Of course, this issue has led to many questions by fellow readers, how a certain portfolio would have performed, if there had been a huge decline in the bond market or if other tail-events would have happened. Since many of these questions have not been answered yet, we would like to analyze and compare ten popular portfolio construction techniques by applying an advanced Monte Carlo simulation to avoid using historical data. Therefore we will be able to get an unbiased view of the pros and cons of each single portfolio construction technique.

 

I: Analyzed Portfolios

In terms of asset selection, we are reviewing following concepts:

 

1. Global Minimum Variance Portfolio (GMV):

This portfolio construction technique has only the objective of lowering risk, rather than aiming to optimize the risk/reward ratio. It creates a portfolio with the lowest possible risk (volatility), which mostly leads to pronounced concentration in low volatility asset classes.

2. Minimum Correlation Portfolio (MCP):

Asset classes with low correlations and volatility relative to other asset classes within the portfolio receive higher weighting. So in the end, the weightings will lead to the effect that all underlying asset classes have the lowest volatility weighted average correlation coefficient to each other. We have used the simple version of the construction technique since the advanced one involves also parts from the Inverse Volatility concept, which is being evaluated separately.

3. Most Diversified Portfolio (MDP):

In contrast to the MCP, the MDP is focused on maximizing the diversification benefits within the portfolio (instead on minimizing the average correlation), by maximizing the diversification ratio, which is defined as the ratio of the portfolio’s weighted average volatility to its overall volatility. Therefore the MDP will utilize the highest degree of diversification benefits. (A broad based MDP is being updated on a regular basis on this website - click here to access the Most Diversified Portfolio).

4. Risk Parity Portfolio (RPP):

The concept is quite straightforward: each asset class should contribute the same amount of risk (volatility) to the overall portfolio. Therefore assets with lower risk, such as bonds, will get a larger part of the portfolio than risky ones.

5. Inverse Volatility Portfolio (IVP):

Each asset is weighted in inverse proportion to its volatility and then all assets are rescaled to sum up to 1. Therefore lower weights are given to high volatility assets and higher weights to low volatility securities. Misleadingly, this concept is often being mixed-up with the risk parity approach, since they are quite similar. Nevertheless, since the overall portfolio volatility is not an additive function of the underlying volatilities, each asset class is not contributing exactly the same amount of risk to the overall portfolio!

6. Minimum Tail Dependent Portfolio (MTP):

Quite unknown so far, this concept is weighting asset classes according to their tail-dependencies. In the first step, it measures the correlation of each asset class during tail events (by using copulas) and then the weights are set to get a minimum tail dependent portfolio.

7. Classic Balanced Portfolio (CBP):

A standard portfolio where 60 percent is being invested in equities and 40 percent in bonds.

8. Momentum Based Portfolio (MBP):

This strategy owns the best two out of four asset classes (equally weighted), which performed the best in the trailing one-month and holds that asset class forward for one additional month.

9. IVY Portfolio (IVP):

The strategy calculates a 10-month moving average for each underlying asset class. If the current price is above its long-term average, the specific asset class will be added within the portfolio otherwise its whole exposure is moved to cash.

10. Permanent Portfolio (PEP):

Harry Browne's "Permanent Portfolio" is investing 25% in stocks, 25% in cash, 25% in gold, and 25% in long-term treasury bonds as a way to cover each of the four economic stages (prosperity, recession, inflation, and deflation).

 

II: Methodology

As already mentioned above, we are using a Monte Carlo simulation, to generate 300 years of daily data. In general, a Monte Carlo simulation performs any kind of analysis by building samples of possible results by substituting a range of values (a probability distribution) for any factor that has inherent uncertainty. It then calculates results over and over, each time using a different set of random values from the probability function. In other words, it is possible to simulate umpteen years of data, which have more or less the same risk characteristics as the desired underlying asset classes.

 

In our example, we will use a multivariate normal distribution (http://en.wikipedia.org/wiki/ Multivariate_normal_distribution), which is fitted with certain parameters (return- and correlation estimates) in order to describe the random behavior (market riskiness) of asset returns. Worth mentioning is the fact that the input parameters are just a point of reference since the multivariate normal distribution will randomly generate values within the normal standard distribution. However, when it comes to simulations, the results should be seen as additional source of information rather than an all explaining framework. For that reason, all the results should be treated with caution and should just be seen as another mosaic stone in quantitative finance.

 

According to a Bridgewater research paper, betas (asset classes) have had, and are expected to have, ratios of excess returns to excess risks (Sharpe Ratios) of about 0.2 to 0.3 (www.bwater.com). That is because:

  • "there needs to be some extra return to compensate investors for excess risk (so this ratio should be positive),
  • but this ratio cannot be very extremely positive because that would make these investments attract substantial amounts of capital that would bid up their prices and lower their expected return."

Therefore we have used an average return estimation for each asset class which produces Sharpe Ratios of 0.2 to 0.3, whereas the risk free rate is 1.5 percent. Furthermore, we have used the historical variance-/covariance matrix from 1980 until 2013 for the iShares Core S&P 500 ETF (IVV), the iShares Core Total U.S. Bond Market ETF (AGG), the UBS ETRACS DJ-UBS Commodity Index Total Return ETN (DJCI) and the SPDR Gold Shares (GLD), as reference point for the underlying correlations assumptions.

 

Summary statistics, drawdowns as well as a performance chart (logarithmic scale) for the four simulated asset classes are given below. As you can see, the Sharpe Ratios for the underlying securities are almost equal, since they are ranging between 0.2 and 0.3. More importantly, with a maximum loss of 87 percent, equities (IVV) have faced the highest decline in the past, followed by gold (GLD), commodities (DJCI) and bonds (AGG). Another interesting fact is the decline of bonds (AGG) by almost 40 percent, a situation which we have not seen in the last couple of decades. Therefore, it will be quite interesting to see, how bond-heavy portfolio construction techniques will perform under such conditions. However, the overall key statistics are looking quite reasonable and therefore the simulated time series can be seen as a quite good proxy for evaluating the above mentioned portfolio concepts.

 

heading_blank_indicator
The Most Rewarding Portfolio Construction Techniques - Performance Ratios
The Most Rewarding Portfolio Construction Techniques - Overview Performance
The Most Rewarding Portfolio Construction Techniques - Drawdowns
heading_blank_indicator

In our example, there is no allowance for transaction costs or brokerage fees. In order to minimize transaction costs, we rebalance the portfolio on a monthly basis, whereas 1-day slippage is included. In addition, we have determined the weights for all portfolio concepts on a monthly basis. Worth mentioning is the fact that we are using a rolling variance/co-variance matrix to determine the weights for all risk-based concepts, as correlations among asset classes are not stable over time.

 

III: Results

a) CAGR

The simulation results are summarized in the table below. The most favorable portfolio in terms of performance is the Minimum Correlation Portfolio (MCP), with an annualized return of 5.87 percent, followed by the Momentum Based Portfolio (MBP) with 5.55 percent and the Most Diversified Portfolio with 5.3 percent. With slightly more than 4 percent, the IVY Portfolio (IVY) is delivering the lowest compound annual growth rate, followed by the Classical Balanced Portfolio (CBP), with an average return of 4.07 percent a year.

heading_blank_indicator
The Most Rewarding Portfolio Construction Techniques - CAGR Performance Ratios
heading_blank_indicator

b) Sharpe Ratios

However, return is just telling one side of the story. For that reason, the Sharpe Ratio is a better measure to examine the risk/reward ratio. The Most Diversified Portfolio (MDP) has the highest Sharpe Ratio, followed by the Permanent Portfolio (PEP) and the Risk Parity Portfolio (RPP). This result is strongly in-line with the theoretical background (https://www.putnam.com/literature/pdf/SU830.pdf), as there should be no other portfolio combination, which can achieve a higher risk-/return ratio than the Most Diversified Portfolio (MDP). In general, 5 out of 10 portfolio concepts had Sharpe Ratios above 0.5. Moreover, the Classic Balanced Portfolio (CBP), as well as all timing based portfolio concepts (MBP, IVY) are delivering the lowest Sharpe Ratios. A quite good reason for that outcome is the fact that those portfolios are not as good diversified as their peers and therefore their overall portfolio volatility is either too high (CBP and the MBP) or their average return is too low (IVY).

c) Volatility/Risk

In terms of risk, the Global Minimum Variance Portfolio (GMP) has of course the lowest portfolio volatility, whereas the Momentum Based Portfolio (MBP) and the Minimum Correlation Portfolio (MCP) are having by far the highest ones.

d) Drawdowns

To evaluate the absolute return character of each portfolio construction technique, we will have a closer look on the ability of each concept to generate absolute returns on a yearly basis as well as on the maximal- and average drawdown statistics of those portfolios.

 

Within almost 300 years of data, the Most Diversified Portfolio (MDP) as well as the Minimum Tail Depended Portfolio (MTP) had only 24 negative years, followed by the Permanent Portfolio (PEP) and the Risk Parity Portfolio (RPR) with 25 years of negative returns. The Global Minimum Variance Portfolio (GMV) failed to deliver positive returns in 27 years, whereas all other portfolio concepts had negative returns in 30 (MBP) years or above (IVY, MCP and the CBP).

 

The IVY Portfolio (IVY) as well as the Global Minimum Variance (GMV) had the lowest maximal drawdown, whereas the Minimum Correlation Portfolio (MCP) as well as the Momentum Based Portfolio (MBP) had by far the highest ones. Another highly interesting fact is that the largest drawdown of most portfolios did not occur within the same time period in which their underlying asset classes suffered the most (see table below).

heading_blank_indicator
The Most Rewarding Portfolio Construction Techniques - Drawdowns
heading_blank_indicator
Especially, the Risk Parity Portfolio (RPP) and the Most Diversified Portfolio (MDP) were holding up quite well, during that time the bond market crashed, while the Global Minimum Variance Portfolio (GMV) was hurt the most! This outcome is not surprisingly at all, since the Global Minimum Variance Portfolio is heavily weighted in bonds and in addition, this concept does not utilize all benefits from diversification. As already mentioned in other articles, as long as the correlations within the underlying asset classes remain low, the Risk Parity Portfolio (RPP) as well as the Most Diversified Portfolio (MDP) won’t be hurt too much, in times of rising interest rates, although they are also heavily weighted in bonds. This is mainly due to the effect, that other asset classes are offsetting the negative performance of bonds. In general, prices do not go down to zero over night and therefore it still makes sense to keep a declining asset class within the portfolio, due to their diversification benefits. For example, if there is a normal correction within an ongoing equity bull market, bonds still tend to rise although they might be in a longer lasting down-trend. This will lead to smoother returns and less drawdowns, which is one of the main strengths of those new portfolio construction concepts. Furthermore, we can see that apart from the Minimum Correlation Portfolio (MCP), the Classic Balanced Portfolio (CBP) and the Global Minimum Variance (GMV), all other concepts had their worst performance more or less during the same time period.

e) Average Losses

Since the historical maximum drawdown is just representing a single tail event in the past, the average drawdowns are much more accurate to evaluate the absolute return character of each construction technique. Below, we have charted all drawdowns as well as the average loss of each portfolio. There we can see that the Permanent Portfolio (PEP) as well as the Most Diversified Portfolio (MDP) had the lowest drawdowns on average, whereas the Minimum Correlation Portfolio (MCP), the Classic Balanced Portfolio (CBP) as well as the Momentum Based Portfolio (MBP) had the highest ones.

heading_blank_indicator
The Most Rewarding Portfolio Construction Techniques - Average Losses
heading_blank_indicator

f) Overall Scoring

To make things more clearly, we have ranked the portfolios with numbers ranging from 10 to 1, whereas 10 represents the most favorable result. Furthermore, we have averaged all scores, to get a total score for each portfolio. There we can see that the Most Diversified Portfolio (MDP) has got the highest average score and is therefore representing a good mixture between performance and risk. The Risk Parity Portfolio (RPP) is on the second place, with a score of 7.5, followed by the Permanent Portfolio (PEP) and Global Minimum Variance Portfolio (GMV). However, with 2 points, the Classic Balanced Portfolio (CBP) has received the lowest score among all evaluated portfolio construction concepts, followed by the Minimum Correlation Portfolio (MCP) and the Momentum Based Portfolio (MBP)!

heading_blank_indicator
The Most Rewarding Portfolio Construction Techniques - Overall Scoring
heading_blank_indicator

IV: Summary

Basically, all portfolio construction techniques are delivering higher Sharpe Ratios than their underlying asset classes, which is just another proof of the overall diversification concept. The Global Minimum Variance Portfolio (GMV) is still doing a good job although it will definitely get hurt the most if we see rising interest rates in the future. As already mentioned in our previous articles, a huge decline in the bond market will not automatically lead to such catastrophic scenarios for the Risk Parity Portfolio (RPP) or the Most Diversified Portfolio (MDP), as some fellow readers might have thought. In contrast, the Inverse Volatility Portfolio (IVP), which is often mixed-up with the risk parity approach, is underperforming its bigger brother (RPP) in terms of its overall score. If we focus on portfolios that are based on timing rather than portfolio construction, we can see that the typical late-in, late-out effect of those concepts will mostly lead to increased volatility or low returns, which is of course affecting their Sharpe Ratios as well as other risk/performance based ratios. The Minimum Correlation Portfolio (MCP) does not utilize all the benefits which can be achieved by diversification and if we have a look at this specific weighting algorithm, we do not really see any big advantage over other concepts. The Minimum Tail Dependent Portfolio (MTP) has shown quite robust results and given a Sharpe Ratio above 0.5, it is definitive worth an investment. However, as this optimization concept needs a lot of estimates (via copula), we think it might be a bit challenging for common investors to understand the rationale behind it. In contrast, the Permanent Portfolio defended its reputation, to be a good and simple investment concept as it has delivered the third highest Sharpe Ratio and the third highest overall score. Furthermore, we can say that the Classic Balanced Portfolio (CBP) is by far the worst concept, followed by the Momentum Based Portfolio (MBP), whereas the Most Diversified Portfolio (MDP) has shown once again that diversification pays off the most, as it is the only free lunch an investor has.

heading_blank_indicator
blank_download_table icon_download_table Download "The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation" as pdf-file
heading_blank_indicator
heading_blank_indicator heading_blank_indicator heading_blank_indicator
heading_blank_indicator
heading_blank_indicator
When will you join our "Model Portfolio Section"?
  • Get a weekly update of our magnificent model portfolios for the ridiculous rate of only a couple of cents a day
  • Plus all the other proven indicators of the Wall Street pros!
heading_blank_indicator
Join the ETF Model Portfolios and Get Instant Access to our Members Section now!
heading_blank_indicator
heading_blank_indicator
heading_blank_indicator heading_blank_indicator heading_blank_indicator
heading_blank_indicator
The Site of Market Technicians, Market Timers & Day Traders

  More information and related content in our ETF Model Portfolio section:

  » WSC All Weather Portfolio! The Perfect Core-Investment For Lazy Investors!
blank blank
  » WSC Inflation Proof Retirement Portfolio! The Perfect Inflation Hedge
blank blank
  » WSC Sector Rotation Strategy! Optimal Downside Protection & High Upside Capture
blank blank
  » WSC Global Tactical ETF Portfolio! The Perfect Strategy For Aggressive Market Timers
blank blank
  » The Most Rewarding Portfolio Construction Techniques: An Unbiased Evaluation! Our research paper!
blank blank
  » Modern Portfolio Theory 2.0 - The Most Diversified Portfolio! Our research paper!
blank blank
  » The Most Diversified Inflation-Proof Retirement Portfolio! Our research paper for investors who need an inflation-proof portfolio!
blank blank

 

heading_blank_indicator
The Site of Market Technicians, Market Timers & Day Traders
heading_blank_indicator
heading_blank_bulletlink « Back to the ETF Model Portfolio Overview page
heading_blank_bulletlink heading_blank_bulletlink
heading_blank_bulletlink « Homepage
heading_blank_indicator
Go to top Top
 
   
blank_1_Technical Market Indicators
Take a Tour - it only takes you 5 minutes.
blank_1_Technical Market Indicators
The Membership Benefits of WallStreetCourier.com
blank_membership_benefits
 checkbox_Proven & Unique Technical Market Indicators  Technical Market Indicators
blank_1 checkbox_The World Famous Smart Money Flow Index The Smart Money Flow Index
blank_1 checkbox_Effective Trend Indicators Effective Trend Indicators
blank_1 checkbox_Powerful Breadth Indicators Powerful Breadth Indicators
blank_1 checkbox_Inimitable Options- & Contrarian Indicators Great Options Indicators
blank_1 checkbox_Inimitable Options- & Contrarian Indicators Inimitable Contrarian Indicators
blank_1 checkbox_Proven Smart- & Dumb Money Indicators Proven Smart Money Indicators
blank_1 checkbox_Proven Smart- & Dumb Money Indicators Proven Dumb Money Indicators
blank_1_Technical Market Indicators blank_1
hor_dotted_Model Portfolios
 checkbox_Model Portfolios  Model Portfolios
blank_1 checkbox_WSC All Weather Model Portfoli WSC All Weather Model Portfolio
blank_1 checkbox_WSC Global ETF Model Portfolio WSC Global ETF Model Portfolio
blank_1 checkbox_Inflation Proof Retirement MP Inflation Proof Retirement MP
blank_1 checkbox_WSC Sector Rotation Strategy WSC Sector Rotation Strategy
blank_1_Model Portfolios blank_1
hor_dotted_Model Portfolios
 checkbox_Heat Maps  Heat Maps
blank_1 checkbox_Global ETF Momentum Heat Map Global ETF Momentum Heat Map
blank_1 checkbox_SPDR Sector Momentum Heat Map SPDR Sector Momentum HM
blank_1_Heat Maps blank_1
hor_dotted_Heat Maps
 checkbox_Weekly Technical Market Comment  Technical Market Comment
blank_1 checkbox_Profitable Long & Short Advice Profitable Long & Short Advice
blank_1 checkbox_Learn From Experts Learn From Experts
blank_1 checkbox_Follow our Trading Advice Follow our Trading Advice
blank_1_Technical Market Comment blank_1
hor_dotted_Technical Market Comment
 checkbox_Dashboards  Dashboards
blank_1 checkbox_Indicator Dashboard Indicator Dashboard
blank_1 checkbox_The CFTC/CoT Dashboard The CFTC/CoT Dashboard
blank_1_Dashboards blank_1
hor_dotted_Dashboards
 checkbox_Fully Packed PDF Slides  Fully Packed PDF Slides
blank_1 checkbox_WSC Technical Market Indicators Paperback (PDF file) WSC Technical Market
blank_1 checkbox_Indicators Paperback Indicators Paperback
blank_1 checkbox_WSC Compendium For The COT Report (PDF file) WSC Compendium: COT Report
blank_1_Fully Packed PDF Slides blank_1
hor_dotted_Fully Packed PDF Slides
 checkbox_Commitments of Traders Report (CFTC/CoT)  Commitments of Traders
blank_1 checkbox_CFTC/CoT Charts for more than 46 markets CFTC/CoT Data for more
blank_1 checkbox_CFTC/CoT Charts for more than 46 markets than 42 markets
blank_1_Commitments of Traders blank_1
hor_dotted_Commitments of Traders
 checkbox_Data Download  Data Download
blank_1 checkbox_Download our entire published indicators Download our entire indicators
blank_1 checkbox_CoT-Data (Commitments of Traders) since 1986 CoT-Data since 1986
blank_1 checkbox_Heat Maps Heat Maps
blank_1 checkbox_Extended Daily Market Statistics Extended Daily Market Statistics
blank_1 checkbox_Weekly Market Statistics Weekly Market Statistics
blank_1 checkbox_Short Interest Data for more than 8,000 U.S. stocks since 2005 Short Interest Data
blank_1_Data Download blank_1
hor_dotted_Data Download
 checkbox_Day Trading Patterns  Day Trading Patterns
blank_1 checkbox_Discover the secrets of the pros Discover the secrets of the pros
blank_1 checkbox_Beat the market by trading only Beat the market by trading only
blank_1 checkbox_on certain market behaviors! on certain market behaviors!
blank_1_Day Trading Patterns blank_1
blank_2
Sign up and become a member of WallStreetCourier.com
blank_1
 
heading_blank_indicator
blank
blank
© 2013 WallStreetCourier.com