Momentum & Value vs. Growth & Value

September 20, 2016

At Dorsey Wright, we believe momentum can be used as a stand-alone investment strategy, however, combining it with other smart beta factors to which momentum is negatively correlated has its advantages.  We have referenced this in previous blog posts, noting that it allows for a portfolio to capture alpha at different periods of the market cycle, which in turn can reduce both drawdowns and volatility.   In this post we would like to discuss the potential benefits of combining momentum with value versus combining growth and value.   Furthermore, we will take a look the correlation of excess returns for each portfolio, and wrap things up by comparing the returns of each.

To begin, let’s take a look at the side by side performance (annual figures) for the products we will be using in our study:  PowerShares DWA Momentum Portfolio PDP, Russell 1000 Growth Index RLG, and the Guggenheim S&P 500 Pure Value ETF RPV.  We can reference this table in comparison to the results we get when combining the smart beta factors we mentioned earlier.  In order to get proper historical data, we used the underlying index (total return) for both RLG and RPV.  For PDP, total return figures were used starting on 3/1/2007.  The table below confirms that when using each of these products as a stand-alone investment product.  As we can see, momentum outperforms all other factors but also at a slightly elevated volatility.   Perhaps the most surprising theme is the underperformance of the growth factor throughout this time frame.

all

PDP inception date: March 1, 2007, RLG inception date: May 22, 2000, RPV inception date:   March 1, 2006 – data prior to inception is based on a back-test of the underlying indexes.   Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividends prior to 3/1/2007.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

Next, let’s take a look at the correlation coefficients when comparing the returns of each portfolio.  Below we’ve plotted the returns of each portfolio against each other on a year-to-year basis.   The correlation of excess returns between PDP and RPV came out to be -.50 during this time period, just slightly better then RLG vs. RPV (which registered -.40).   Again, both of these are impressive in terms of negative correlation which hopefully will give us the ability to capture alpha at different areas of the market cycle once we construct our portfolios.   Typcially our goal in doing this is lowering portfolio volatility and reducing max drawdowns when compared to using them as stand alone investments.

pdp-vs-rpv

PDP inception date: March 1, 2007, RLG inception date: May 22, 2000, RPV inception date:   March 1, 2006 – data prior to inception is based on a back-test of the underlying indexes.   Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividend prior to 3/1/2007.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

rlg-vs-rpv

PDP inception date: March 1, 2007, RLG inception date: May 22, 2000, RPV inception date:   March 1, 2006 – data prior to inception is based on a back-test of the underlying indexes.   Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividend prior to 3/1/2007.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  

In conclusion, the portfolios we will construct are going to be based on a static allocation of 70%/30%.  To clarify, both the momentum and growth allocations will remain at 70%, while the value portion will be 30%.  The portfolios are re-balanced annually (although as we mentioned the allocation will remain static).    Looking at the table below, we can see that the momentum/value combination portfolio outperformed has over the growth/value combination.   The returns are nearly double, while volatility remains the same at 22%.   Market participants looking to combine a portion of their value portfolio with another allocation would certainly seem to benefit by using a momentum product vs. a growth product.

summary

PDP inception date: March 1, 2007, RLG inception date: May 22, 2000, RPV inception date:   March 1, 2006 – data prior to inception is based on a back-test of the underlying indexes.   Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividend prior to 3/1/2007.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

DWA provides the underlying index for PDP (discussed above) and receives licensing fees from Invesco PowerShares based on assets invested in the Fund.

Some information presented is the result of a strategy back-test.  Back-tests are hypothetical (they do not reflect trading in actual accounts) and are provided for informational purposes to illustrate the effects of the strategy during a specific period.  Back-tested results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money.  Back-testing also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight.  

Neither the information within this email, nor any opinion expressed shall constitute an offer to sell or a solicitation or an offer to buy any securities, commodities or exchange traded products.  This article does not purport to be complete description of the securities or commodities, markets or developments to which reference is made. 

The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  Relative Strength is a measure of price momentum based on historical price activity.  Relative Strength is not predictive and there is no assurance that forecasts based on relative strength can be relied upon.

 

Posted by:


Weekly RS Recap

September 6, 2016

The table below shows the performance of a universe of mid and large cap U.S. equities, broken down by relative strength decile and quartile and then compared to the universe return.  Those at the top of the ranks are those stocks which have the best intermediate-term relative strength.  Relative strength strategies buy securities that have strong intermediate-term relative strength and hold them as long as they remain strong.

Last week’s performance (8/29/16 – 9/2/16) is as follows:

ranks 09.06.16

This example is presented for illustrative purposes only and does not represent a past or present recommendation.  The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.

Posted by:


Quote of the Week

August 31, 2016

Leda Braga via Financial News:

The world is learning the value of data-driven activities through science and technology. I can imagine in a few years’ time calling a self-driven cab – it will be reliable and cost-effective. Investment management is one of the most data-driven activities there are. It is the perfect ground for the systematic approach.

HT: Jerry Parker/Twitter

Posted by:


Weekly RS Recap

August 29, 2016

The table below shows the performance of a universe of mid and large cap U.S. equities, broken down by relative strength decile and quartile and then compared to the universe return.  Those at the top of the ranks are those stocks which have the best intermediate-term relative strength.  Relative strength strategies buy securities that have strong intermediate-term relative strength and hold them as long as they remain strong.

Last week’s performance (8/22/16 – 8/26/16) is as follows:

ranks

This example is presented for illustrative purposes only and does not represent a past or present recommendation.  The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.

Posted by:


Weekly RS Recap

August 15, 2016

The table below shows the performance of a universe of mid and large cap U.S. equities, broken down by relative strength decile and quartile and then compared to the universe return.  Those at the top of the ranks are those stocks which have the best intermediate-term relative strength.  Relative strength strategies buy securities that have strong intermediate-term relative strength and hold them as long as they remain strong.

Last week’s performance (8/8/16 – 8/12/16) is as follows:

ranks

This example is presented for illustrative purposes only and does not represent a past or present recommendation.  The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.

Posted by:


Weekly RS Recap

July 25, 2016

The table below shows the performance of a universe of mid and large cap U.S. equities, broken down by relative strength decile and quartile and then compared to the universe return.  Those at the top of the ranks are those stocks which have the best intermediate-term relative strength.  Relative strength strategies buy securities that have strong intermediate-term relative strength and hold them as long as they remain strong.

Last week’s performance (7/18/16 – 7/22/16) is as follows:

ranks

Performance by sector for the week is shown below:

sector ranks

This example is presented for illustrative purposes only and does not represent a past or present recommendation.  The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.

Posted by:


Sector Performance

July 22, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 7/21/16.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Combining Equity Momentum (PDP) & Managed Futures (CSAIX)

July 11, 2016

One of the more surprising themes for many market participants thus far into 2016 has been the impressive strength of the commodities asset class.   What started off as a sharp rally from the lows in the energy sector quickly flowed into other commodities such as precious metals, softs (sugar) and ags (soybeans).   A lack of trend in US equities has helped funds flow into commodities as market participants search for yield in other asset classes.   Another tailwind for commodities has been the range bound activity in the US Dollar.  This in turn makes them cheaper to export overseas and therefore increases demand.

Although it may not be the most popular asset class amongst traditional fund managers, there is certainly demand for commodities markets among institutional traders.   For example, managed money (CTA’s and Macro Hedge Funds) is often focused on looking for trending markets in both foreign exchange and commodities to allocate toward their portfolios.  Given the size and leverage these funds have access too, the commodities and foreign exchange markets are often subject to large trending moves which can continue for extended periods of time.  Although momentum and trend following are often used interchangeably, they actually differ in the fact momentum (ex. 12 mo. trailing return) is typically thought of as relative while trend following techniques (ex. moving averages) are more absolute in nature.  Our main point in mentioning CTA’s and Hedge Funds is that given their ability to rotate through various asset classes and take both sides of the market (long or short) they typically have a negative correlation to long only equity managers.

Assuming the negative correlation between the two strategies holds true, combining a long only equity momentum portfolio with some type of managed futures strategy would certainly seem make sense in terms of reducing volatility and drawdowns while maintaining alpha above a related benchmark.  Let’s investigate this matter further by using the Power Shares DWA Momentum Portfolio (PDP) and combining it into a portfolio with the Credit Suisse Managed Futures Strategy Fund (CSAIX).   Note we will be using the returns of the underlying index (CSTHFMF0 – Credit Suisse Hedge Fund Index Managed Futures) in order to pull the historical data for CSAIX since its inception was 9/28/12.

Here is a brief summary of each strategy side by side.     As shown below, PDP outperforms both the Credit Suisse Hedge Fund Index Managed Futures Strategy CSAIX and SPX over the allotted time frame in this study.  However, as we saw in our previous posts it does so with slightly elevated volatility.   All other performance metrics aside, the main concept we want to emphasize is the differences in returns each year between PDP (momentum) and CSAIX (managed futures).  The most obvious example is 2008, when CSAIX posted an impressive 18.33% gain while both PDP and the SPX suffered steep double digit losses.

Click on graphic for larger version

ANNUAL RETURNS

PDP inception date: March 1, 2007, CSAIX inception date: Sept 28, 2012 – data prior to inception is based on a back-test of the underlying indexes Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

The graphic below shows a comparison of annual returns using the time period 1998 and 2015 in which the correlation of excess returns between momentum and managed futures  was  (-0.91)  A few of the outlier years to take note of which contain major differences in performance are 1999, 2008, 2009, and finally 2014.   Some of the largest differences in performance can be attributed to periods of heighted equity market volatility (ex 2008).   Excess volatility tends to create more opportunities for managed futures strategies.  On the other hand, the past 5 years (with the exception of 2014) showed equity momentum outperforming managed futures as the stock market continued its strong bull market while many commodities and foreign exchange rates were lacking volatility and any type of sustained trend (up or down).

Click on graphic for larger version

correlations

PDP inception date: March 1, 2007, CSAIX inception date: Sept 28, 2012 – data prior to inception is based on a back-test of the underlying indexes Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

The table below goes through similar allocation structure of how we set up our low volatility and value portfolios in combination with momentum in previous posts.  Our main goal here is to stay consistent and create a robust process that has very few moving parts.  The portfolios are all re-balanced annually but each allocation remains consistent each year.   The main goal is to emphasize the benefit of combining two negatively correlated strategies in order to take advantage of the performance differences each will achieve throughout different market cycles.

Click on graphic for larger version

HISTORICAL ALLOCATIONS

The returns above are based on hypothetical back-tests of the various allocation options.  PDP inception date: March 1, 2007, CSAIX (inception date: Sept 28, 2012 – data prior to inception is based on a back-test of the underlying indexes. Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision-making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight.  PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

Let’s take it a step further and add one simple risk metric to our portfolio and see if we can reduce volatility even further.   Over time, it’s typically been the case that a long only equity momentum/trend following based strategy tends to perform better while the SPX is above its 200 day moving average.   On the flip side, when the SPX is below its 200 day moving average, periods of heightened volatility are more frequent and can lead to steep draw downs in these portfolios.   This is not always the case but over the years research has shown that the 200 day moving average is often considered a reliable proxy for a risk on/risk off environment.

Interestingly enough, this “risk off” environment is often where managed futures strategies thrive and tend to see their best results.  One main reason for this is the abundance of “fat tail” trades that seem to occur during these market cycles.  The below table compares a model we have created using a 200 day moving average as a risk proxy to determine how we will allocate our portfolio using equity momentum (PDP) and a managed futures strategy (CSAIX).   The allocation will be 80% equity momentum/20% managed futures when the S&P is above its 200 day moving average and 80% managed futures/20% equity momentum when it is below.  In order to reduce turnover, the portfolio will only be re-balanced on a month-end basis.

Click on graphic for larger version

CSAIX

PDP inception date: March 1, 2007, CSAIX  inception date: Sept 28, 2012 – data prior to inception is based on a back-test of the underlying indexes Performance of the switching strategy is the result of back-testing.  Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision-making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight. PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

Let’s take it a step further and add one simple risk metric to our portfolio and see if we can reduce volatility even further.   Over time, it’s typically been the case that a long only equity momentum/trend following based strategy tends to perform better while the SPX is above its 200 day moving average.   On the flip side, when the SPX is below its 200 day moving average, periods of heightened volatility are more frequent and can lead to steep draw downs in these portfolios.   This is not always the case but over the years research has shown that the 200 day moving average is often considered a reliable proxy for a risk on/risk off environment.

Interestingly enough, this “risk off” environment is often where managed futures strategies thrive and tend to see their best results.  One main reason for this is the abundance of “fat tail” trades that seem to occur during these market cycles.  The below table compares a model we have created using a 200 day moving average as a risk proxy to determine how we will allocate our portfolio using equity momentum (PDP) and a managed futures strategy (CSAIX).   The allocation will be 80% momentum/20% managed futures when the S&P is above its 200 day moving average and 20% equity/80% managed futures when it is below.  In order to reduce turnover, the portfolio will only be re-balanced on a month-end basis.

 

 

Posted by:


Weekly RS Recap

June 27, 2016

The table below shows the performance of a universe of mid and large cap U.S. equities, broken down by relative strength decile and quartile and then compared to the universe return.  Those at the top of the ranks are those stocks which have the best intermediate-term relative strength.  Relative strength strategies buy securities that have strong intermediate-term relative strength and hold them as long as they remain strong.

Last week’s performance (6/20/16 – 6/24/16) is as follows:

ranks

This example is presented for illustrative purposes only and does not represent a past or present recommendation.  The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.

Posted by:


A Momentum Based Core Equity Strategy (Part 4)

June 14, 2016

To read Part 1 click here

To read Part 2 click here

To read Part 3 click here

The Core Equity strategy uses the three factor strategies discusses in Part 3.  In order to keep things as simple and non-optimized as possible I just included each model at a 1/3 weight of the total portfolio.  However, by using three different models you can adjust the weightings of each one to suit the end investor’s needs.  For example, if you needed less volatility you can just increase the weight of that model in the combined portfolio.

The Core Equity strategy is rebalanced monthly just like the individual factor models are.  Since a stock can be included in more than one of the factor models, the final model isn’t necessarily 300 stocks with equal weights.  Some stocks will have larger weights than others because they are in multiple models.  That is the only weighting difference though – there is no market capitalization or factor adjustment made to the stock weightings.

CE

In order to judge the fourth factor, size, I modified the final portfolio construction process to account for market cap.  Each stock’s market cap along with how many of the three factor models it was in was taken into account.  A mega cap stock in only one of the factor models might have a higher weight than a smaller cap stock in multiple models in this scenario.  Generally speaking, market capitalization weighting is sub-optimal for investment returns (not for capacity) and that is one big reason why Smart Beta has taken off.  We see the same thing here when we adjust for market cap.

CE2

The returns for the final Core Equity strategy add significant value over the broad market while keeping the volatility (standard deviation) close to that of the benchmark (below for a cap weight version of the Core Equity strategy).  More importantly, it helps smooth out the ride of the individual factor models.  In 1999, for example, Low Volatility was a large underperformer, but momentum was strong enough to carry the overall portfolio to strong returns.  Just two years later in 2001, the roles were reversed and it was Low Volatility and Value picking up the slack for the poor momentum returns.  You can see similar things happening in years like 2006, 2007, and 2011.  I think this point is vastly underrated because investors tend to abandon strategies at exactly the wrong times – after they have underperformed and are due for a rebound.  Combining all three factor strategies into one large Core Equity portfolio helps mask the underperformance of specific factor models and helps investors stay with underperforming strategies.

There are probably an infinite number of ways you can construct a core equity strategy using different factors.  Here, I have looked at just one that was designed to be extremely simple, non-optimized, and robust going forward.  I used some custom models to create the underlying factor models, but they shouldn’t be so dramatically different from existing ETF’s or mutual funds that something similar couldn’t be done on a smaller scale.  I’m sure there are better ways to construct the factor models and put them together in the final Core Equity strategy.  If you have any ideas about how that can be done I would love to hear them!

 

The returns used within this article are the result of a back-test using indexes that are not available for direct investment.  Returns do include dividends, but do not include transaction costs.  Back-tested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes to illustrate the effects of the discussed strategy during a specific period.  Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight.  Dorsey, Wright & Associates believes the data used in the testing to be from credible, reliable sources, however; Dorsey, Wright & Associates, LLC (collectively with its affiliates and parent company, “DWA”) makes no representation or warranties of any kind as to the accuracy of such data. Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  

Posted by:


A Momentum Based Core Equity Strategy (Part 3)

June 13, 2016

To read Part 1 click here

To read Part 2 click here

I am using three different factor strategies (Momentum, Low Volatility, and Value) to form a Core Equity strategy.  A fourth factor, Size, will also be considered in the portfolio construction process (equal weighted instead of cap weighted).

All three of the factor strategies are constructed in a similar way.  The universe is the top 1000 market cap stocks traded in the U.S.  All of these stocks should have plenty of liquidity and should eliminate a lot of the issues that occur when you are dealing with small or micro-cap stocks.  All three strategies are rebalanced at the end of each month and have 100 stocks.  The strategies are run separately so in theory a stock could appear in all three strategies (I will account for this in the final model).  In addition, all three strategies use a Point and Figure momentum overlay that was discussed in part 2.  One of the goals was to make the three factor strategies as similar as possible in terms of portfolio construction to avoid as much optimization and curve fitting as possible.  All three of the factor strategies use very simple metrics and I believe they should be robust going forward.  There will certainly be times of underperformance (sometimes dramatic underperformance) from each of the strategies, but over time they have all shown to work very well.

The momentum strategy uses a simple, well-known momentum measure with a Point and Figure relative strength overlay.  Each month stocks are ranked by their trailing 250 day performance skipping the most recent 20 days (or one month).  This is a pretty standard definition for momentum.  The Point and Figure overlay actually does help returns over time, but not anywhere near what it does for the Value and Low Volatility models.  However, I wanted to include the Point and Figure overlay to keep the momentum model similar in portfolio construction to the other two.  Momentum provides really good returns, but is really volatile.  Momentum also has the distinction of being the least correlated with the other factors so they really help smooth out the volatility of a stand-alone momentum strategy.

Mom

The Value strategy uses a composite of four value ratios to rank the stocks in the universe: Price/Sales, Price/Book, Price/Free Cash Flow, and Price/Earnings.  Again, we use a Point and Figure relative strength overlay to improve the returns and filter out some of the value traps.  The universe is the same as the momentum model as are the rebalance dates and weightings.

ValMom

The Low Volatility strategy uses the standard deviation of daily returns over the trailing year to rank stocks in the universe.  The Point and Figure overlay is also used to keep everything consistent, and the other portfolio construction parameters remain the same.

LVMom

All three models perform very well versus the broad market (S&P 500 Total Return).  They do have their bumps along the way in terms of when they outperform and underperform, but often one strategy’s underperformance is offset by outperformance in another.

FactorRet

 

The returns used within this article are the result of a back-test using indexes that are not available for direct investment.  Returns do include dividends, but do not include transaction costs.  Back-tested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes to illustrate the effects of the discussed strategy during a specific period.  Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight.  Dorsey, Wright & Associates believes the data used in the testing to be from credible, reliable sources, however; Dorsey, Wright & Associates, LLC (collectively with its affiliates and parent company, “DWA”) makes no representation or warranties of any kind as to the accuracy of such data. Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  

Posted by:


A Momentum Based Core Equity Strategy (Part 2)

June 10, 2016

To read Part 1 click here

The first post in this series laid out the background for a Core Equity strategy.  The goal is to use different factor strategies to create a more efficient portfolio than what you would get from traditional cap weighting.

The Value and Low Volatility strategies actually have a momentum component to them.  As I mentioned in the background post, we work with momentum all the time so it is easy for me to incorporate that with other factors to try and improve returns.  I realize this will lower some of the correlation benefits, but I think the potential return trade-off is worth it.

The graph below shows two versions of a Value model.  The Pure Value model selects the 100 cheapest stocks based on a composite indicator of Price/Sales, Price/Book, Price/Free Cash Flow, and Price/Earnings.  The Pure Value model below rebalances the portfolio quarterly with the 100 cheapest stocks out of a universe of the top 1000 market cap names.  The ValMom model uses the same value composite ranking system, but requires the stocks to be on a Point and Figure Buy Signal plus be in a Column of X’s.  For those of you not familiar with Point and Figure relative strength, this simply means the stock has been outperforming the broad market on an intermediate and long-term basis.

Val

You can see adding the momentum overlay to the value strategy is beneficial to returns.  Essentially, this helps filter a lot of the value traps.  More experienced Value investors have other methods to accomplish this goal, but adding the momentum overlay is something that works very well for what we do.

I did the same thing with the Low Volatility model.  This model picked 100 stocks with the lowest trailing one year daily standard deviation from a universe of 1000 top market cap names.  The portfolio was rebalanced quarterly.  Adding a momentum overlay helps ensure you have a portfolio of stocks that isn’t volatile, but has also demonstrated the ability to outperform the broad market over time.  Again, this will cut in to some of the correlation benefits of the factor strategies, but I think the return tradeoff is worth it.

LowVol

Adding the momentum overlay doesn’t improve returns as much as the Value example above, but it definitely does help.

The next post will detail the actual models I used for the factor strategies inside of the Core Equity model, but what has been discussed above is the basis for why I’m doing a couple of things differently than other models you might see and why the final model will have more of a momentum tilt.

 

The returns used within this article are the result of a back-test using indexes that are not available for direct investment.  Returns do include dividends, but do not include transaction costs.  Back-tested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes to illustrate the effects of the discussed strategy during a specific period.  Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight.  Dorsey, Wright & Associates believes the data used in the testing to be from credible, reliable sources, however; Dorsey, Wright & Associates, LLC (collectively with its affiliates and parent company, “DWA”) makes no representation or warranties of any kind as to the accuracy of such data. Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  

Posted by:


A Momentum Based Core Equity Strategy (Part 1)

June 9, 2016

Smart Beta strategies have become increasingly popular over the last few years.  These factor based strategies rank groups of stocks or certain characteristics that have proven to provide long-term outperformance over broad market benchmarks.  The strategies tend to be extremely disciplined in their stock selection processes, which has been one of the biggest knocks on active management over the years.

One of the best ways to use the idea of Smart Beta or Factor Investing is to combine them to form a core equity portfolio.  Individually, the factor portfolios often have greater volatility than the overall market.  But many of the factors have excess returns that are negatively correlated so you can combine individual factors together to lower the volatility of the portfolio.  The lower volatility can come with a nice benefit: the factor portfolios have the potential to generate better returns than the overall market.

I was working on another project that involved putting some factor data together so I decided to combine the portfolios into a Core Equity portfolio to see how it looked.  I’ll have a few posts on the blog that run through the steps and the data I used to create the factor strategies as well as the combined strategy.  Since I have developed a lot of momentum strategies over the years you will notice that most of what is going in to this model has a momentum bias.  I’m sure there are some other ways to go about this, but I have worked with the momentum factor for so long it is really easy for me to incorporate it into other factor based strategies.  I thought it would be a useful exercise to run through this process on the blog because we generally get some interesting feedback and suggestions on how to improve the processes.

There are quite a few factors out there, but we will really focus on a manageable number of four: Momentum, Low Volatility, Value, and Size.  I took care of the Size factor in the portfolio construction process.  All of the strategies are equally weighted, which gives them a small cap tilt over time.  I will revisit the effect of the equal weighting tilt at the end when I look at the combined portfolio.  It is easy to estimate the size effect by running a cap weight and an equal weight version of the same portfolio.

The other three factors really form the backbone of the core equity strategy.  In the next post I will discuss how those strategies where constructed and the historical performance of each factor portfolio.

Posted by:


Sector Performance

June 3, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 6/2/16.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Factor Investing: The Benefits of Combining Momentum & Value

June 2, 2016

After our previous write up regarding the idea of combining momentum and low volatility into a portfolio, we had a few requests asking about the concept of combining momentum and value. As long time Dorsey Wright readers know, while we absolutely believe momentum can work as a stand-along strategy in a portfolio, combining momentum with a value based strategy does have certain advantages.   As is the case with low volatility, a value based approach can also typically be thought of as a reversion to the mean type of trade as market participants seek value in underperforming stocks or asset classes.   Obviously, a momentum based approach is focused on finding stocks that have outperformed their peers over a certain period (ex. 12 month trailing), hoping those strong trends continues to maintain leadership in the market.

Here is a brief summary of each strategy side by side.   Note we are using total return for the RPV (Pure Value) and SPX (Benchmark).  As shown below, PDP (Momentum) outperforms both RPV and SPX over the allotted time frame in this study, but not without slightly higher volatility.   We can also see a number of years in which momentum and value have substantially different performance numbers.   Let’s take this a step further and dive into some of the details on the correlation of excess returns between the two strategies.

PDP

As we pointed out in our previous post, the correlation of excess returns between low volatility and momentum came in at roughly -.70.  Given what we mentioned about the underlying theme of momentum investing (trend following) while compared to value investing (mean reversion), it’s logical to think a similar type of figure would exist between these two strategies as well.   The table below shows a comparison of annual returns using the time period 1998 and 2015 in which the correlation of excess returns between value and momentum comes out to be -.50.  A few of the outlier years to take note of which contain major differences in performance are 1998 – 2002, 2007, 2009, and finally 2015.  The main concept again being not only does having the ability to rotate or combine these two factor based strategies help improve performance; it also helps in reducing volatility. (click on below graphic to enlarge)

PDPVSRPV CORR

PDP inception date: March 1, 2007, RPV inception date: March 3, 2006 – data prior to inception is based on a back-test of the underlying indexes. Please see the disclosures for important information regarding back-testing.  PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

Let’s take this a step further to answer the one question that is usually on most market participants minds.   “How do I implement these products into a portfolio for my clients?”   The number of possibilities is endless but to keep things simple we set up a “static allocation” model that rotates through a number of different portfolio’s starting with a 90% PDP/10% RPV and all the way to 10% RPV/90% PDP during our time period stated above.   This gives a very detailed description of how each one of these portfolios would have performed on a cumulative, annual, and risk –adjusted (volatility) basis.   However, what if we could improve on these returns and be more flexible in our allocations.   We know that often times combing some sort of trend following proxy (typically a moving average) in addition to a stand-alone momentum strategy can often times help improve these numbers.    This will be part of our final discussion. (click on below graphic to enlarge)

pdprpvhistorcials

Over time, it’s typically been the case that a momentum/trend following based strategy (assuming a long only portfolio) tends to perform better while the SPX is above its 200 day moving average.   On the flip side, a choppy market with a lack of sustained leadership (which can favor a value based strategy) is more likely to presents itself when the SPX is below its 200 day MA.  This is not always the case but over the years research has shown that the 200 day moving average is often considered a reliable proxy for a risk on/risk off environment.  The below table compares a model we have created using a 200 day moving average as a risk proxy to determine whether or not we will invest in a momentum (PDP) or value (RPV).   We thought it would be interesting to take these allocations to the extreme, relying on a 100% momentum based strategy when the SPX is above its 200 day MA, while flipping to Value when the SPX is below its 200 day moving average.   Our momentum/trend following model which incorporated the 200 day risk proxy averaged just over 10% annual return, while minimizing volatility to just 21%.   Comparing these towards using PDP or RPV as stand-alone vehicles as was in table 1, we can see the benefits when it comes to having access to both products. (click on below graphic to enlarge)

200ma NEW

PDP inception date: March 1, 2007, RPV inception date: March 3, 2006 – data prior to inception is based on a back-test of the underlying indexes. Performance of the switching strategy is the result of back-testing.  Back-tested performance results have certain limitations.  Such results do not represent the impact of material economic and market factors might have on an investment advisor’s decision-making process if the advisor were actually managing client money.  Back-testing performance also differs from actual performance because it is achieved through retroactive application of a model investment methodology designed with the benefit of hindsight. PDP returns do not include dividends.  Returns do not include all potential transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss. 

As this paper shows, there are a number of ways to combine both momentum and value into a portfolio.   For those investors willing to accept slightly higher volatility to achieve higher returns, a portfolio with a larger allocation towards momentum certainly is favorable.   The same can be true for those investors looking to low annualized volatility who might not be as concerned about achieving a certain level of excess return.   Finally, we showed adding a trend following proxy (the 200 day moving average) can help aide in substantial performance over the benchmark (SPX), and also help achieve better risk management when using a momentum or value based strategy as a stand-alone vehicle.

Performance data for the model is the result of hypothetical back-testing.  Performance data for RPV prior to 03/01/06 and PDP prior to 3/01/2007 is the result of backtested underlying index data.  Investors cannot invest directly in an index, like the SPX.  Indexes have no fees.  Total return figures are used in RPV and SPX calculations.  Back-tested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes to illustrate the effects of the strategy during a specific period.  Back-tested performance results have certain limitations. Back-testing performance differs from actual performance because it is achieved through retroactive application of an investment methodology designed with the benefit of hindsight. Model performance data as well as back-tested performance do not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money. Past performance is not indicative of future results. Potential for profits is accompanied by possibility of loss.

Posted by:


Sector Performance

May 6, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 5/5/2016.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Combining Momentum & Low Volatility for Enhanced Alpha

May 2, 2016

Most market participants would agree that four of the most popular factor based investment methods used today are often considered to be momentum, value, growth, and low volatility.   At Dorsey Wright, we are often asked the best way to combine Dorsey Wright strategies (momentum/relative strength) with these other commonly used factor strategies.      Proponents of any smart beta strategy will often support all of these strategies, even admitting that there are pros and cons to each factor.  Momentum, for example, is the idea of investing in securities or asset classes using a previous time period that has performed well, most commonly a 12 month trailing return.  This type of strategy tends to do well during periods of sustained trends, but lags others such as value and low volatility during choppy markets.

During the 1st quarter of 2015, the majority of momentum/relative strength based strategies tended to fare better than the other factor based strategies mentioned above.  One of the largest contributing factors to this alpha generation during Q1 of 2015 was the dispersion which existed amongst US equities.   For example, the energy sector saw a sharp decline while sectors such as healthcare, biotech, and consumer discretionary fared much better.    Fast forward to the Q1 of 2016 and we have a different story on our hands.   Momentum strategies have struggled due to a lack of sustained sector leadership, while investment themes such as low volatility and value have performed much better.

Given the recent changes in sector leadership, we thought it would be interesting to go back and take a look how a few of these factor methods have compared to each other in terms of performance, volatility, etc.  The table below is a simple study complied over the last 18 years comparing PDP (Momentum), SPLV (low volatility), and SPY (benchmark).   Although momentum (PDP) outperforms both SPLV (low volatility) and SPY (benchmark), the added alpha generation came with a few drawbacks (mainly the potential for elevated volatility).   While periods such as these can be difficult, there are certainly ways to minimize the downside when they do come about in the market.   For example, using a systematic process can be a huge advantage, as it helps remove the human emotion which often times is magnified during periods of heightened market volatility.   Let’s take a look at the table below and see what type of results each of these portfolios (momentum, low volatility, and the equity index) generated over the allotted time period.

pdp

PDP inception date: March 1, 2007 – data prior to inception is based on a back-test of the underlying index.

SPLV inception date: May 5, 2011 – data prior to inception is based on a back-test of the underlying index.

When we take a closer look at the returns, we can see just how much the momentum and low volatility factors  differ in terms of performance at various cycles in the market (note 1999, 2003, and 2008 just to name a few).   The idea of implementing both momentum and low volatility into a portfolio would then sure seem logical to most money managers.   After all, any type of low volatility factor investing can typically be thought of as a reversion to the mean type of trade, which most would agree is the exact opposite goal of momentum investing (i.e.  looking for “fat tail” trades that deviate from the mean).    More simply stated, combining two different factor allocations in a portfolio which tend to do well during different market cycles would certainly seem to be an added benefit for any portfolio manager looking to reduce volatility and continue to generate alpha.

pdpsplv

The graphic above does a good job of displaying the differences returns year in and year out.  In fact, the correlation of excess returns between momentum and low volatility ends up at roughly-.70.    We plan to further visit this topic in our next blog post in order to give readers an a better idea on the type of results seen when combining these two factors in both a static and flexible allocations.   For now, the important thing to keep in mind is that in today’s investment world market participants should take full advantage of the full suite of products out there in order to help achieve alpha for their clients.    Using momentum and low volatility is just one way this can be done.   More detailed performance and risk analysis to follow in our next post on this topic.

Performance data for SPLV prior to 05/05/2011 and PDP prior to 3/01/2007 is the result of backtested underlying index data.  Investors cannot invest directly in an index.  Indexes have no fees.  The returns of the ETFs above do not include dividends, or all transaction costs.  Back-tested performance is hypothetical (it does not reflect trading in actual accounts) and is provided for informational purposes to illustrate the effects of the strategy during a specific period.  Back-tested performance results have certain limitations. Back-testing performance differs from actual performance because it is achieved through retroactive application of an investment methodology designed with the benefit of hindsight. Back-tested performance does not represent the impact of material economic and market factors might have on an investment advisor’s decision making process if the advisor were actually managing client money. Past performance is not indicative of future results. Potential for profits is accompanied by possibility of loss.

Neither the information within this post, nor any opinion expressed shall constitute an offer to sell or a solicitation or an offer to buy any securities  This article does not purport to be complete description of the securities to which reference is made.

DWA provides the underlying index for the PDP, discussed above, and receives licensing fees from PowerShares.

           

Posted by:


Sector Performance

April 28, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 4/27/16.

gics

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Performance Within Context of Expectations

April 25, 2016

Just how baffling can the stock market be to investors?  So often the market just does not behave the way investors think it should.   Morgan Housel of The Motley Fool provides some data points that can make investor’s heads explode:

Coca-Cola is fighting 12 consecutive years of soda consumption decline. Its stock is at an all-time high.

Tesla is changing the world, and orders for its new car are off the charts. Its stock is lower than it was 18 months ago.

Cigarette consumption has dropped 44% since 1981. Altria stock is up 71,000% since 1981.

WalMart net income has tripled since 2000. Its stock has lost 1.5% since 2000.

Apple has earned almost a quarter trillion dollars of profit since 2012. Its stock has barely budged.

Amazon’s profits round to zero since 2012. Its stock has tripled.

2009 was one of the worst years for the economy in a century. The market rose 27%.

2015 was a good year for the economy. The market rose 1%.

Brazil’s economy is a disaster. Its stock market is flat over the last two years.

America is enjoying the longest streak of low unemployment claims in four decades. Its stock market is also flat over the last two years.

And so on.

Housel sums up the problem:

Outcomes are determined by performance within the context of expectations, with importance heavily weighted toward the latter. And if predicting future performance is hard, calibrating them against expectations is close to sorcery…

…In a world where analysts focus most of their time analyzing performance – what earnings will do, or what the economy will do – and it’s no wonder we struggle to predict outcomes.

This is where many investors simply throw up their hands and give up on finding a logical, organized way to analyze the market.  It is also where investors who are introduced to the Point and Figure method of technical analysis “see the light” in the sense that the market gets boiled down to understanding that price is the intersection between supply and demand.  The motivation for buying and selling activity may remain elusive, but the imbalance between supply and demand can be seen in the chart.  Momentum of the trend of the security can be identified and derived by looking at the relative strength of the security compared to all other securities in the investment universe.  With that information, investors can invest in the market as it really is and not as they wish it to be.

Posted by:


Sector Performance

April 15, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 4/14/16.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Most investing is simple, but we complicate it.

April 6, 2016

Morgan Housel, in a 2014 WSJ article, shared some wise words to help demystify the stock market.

Companies earn a profit.  When investors are in a good mood, they pay up for that profit.  When they are in a bad mood, they pay less.  Future stock returns will equal profit growth, plus or minus the change in investor attitudes.

That really is all that is going on in the stock market.  But we complicate it, scrutinizing every market detail for evidence of what is coming next.

At their core, market forecasts are an attempt to predict investor’s emotions—say, how happy people will be in 2024.  An there is just no reliable way to do that.

A sensible way to invest is to assume companies will earn a profit, and assume the amount investors will be willing to pay for that profit will fluctuate sporadically.  Those emotional swings will balance out over time, and over the long run the profits companies earned will accrue to investor’s pockets.

Everything else—what stocks might do next quarter, or when the next crash might come—can be needlessly complicating.  Investors should learn to take the simple route.

Housel’s description of how the stock market works is spot on.  However, investors still have to decide what to do with this information.  Perhaps, they will choose buy and hold index investments.  Perhaps, they will choose to employ active investment strategies in an attempt to improve the risk/return profile of a passive investment.  If they choose the latter, even for a portion of their overall allocation, they would be well served to choose active investment strategies that take into account the reality that stock prices are determined by a combination of factors that include corporate profits AND investor emotions.

Key to the rationale for momentum investing is that one never knows, nor does one necessarily care, the exact motivation of buyers and sellers in the marketplace.  For momentum investors, it is enough to see the net effect of all buying and selling pressure for stocks and then to rank a universe of securities by their momentum, buying the securities with the strongest momentum and holding them for as long as they remain strong.

What should investors avoid?  As Housel points out, forecasting is an exercise in futility.  Furthermore, expecting the stock market to be a simple math problem related to corporate profits is another way for investors to set themselves up for disappointment.

The relative strength strategy is NOT a guarantee.  There may be times where all investments and strategies are unfavorable and depreciate in value.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  

Posted by:


Corporate Profits and Stock Market Returns—It’s Complicated

April 4, 2016

There is never any shortage of big scary things to worry about when it comes to the stock market.  Savita Subramanian from Bank of America recently weighed in on her current concerns on earnings growth:

Publicly traded companies have seen negative earnings growth two quarters in a row and there are no fundamental underpinnings for the rally, Savita Subramanian, BofAML’s head of U.S. equity and quantitative strategy, said on CNBC’s “Fast Money” this week.

“We are in a profits recession. There (are) no two ways around it,” said Subramanian, whose S&P 500 price target of 2,000 is among the lowest on Wall Street. She is also concerned about how Federal Reserve monetary policy could affect stocks.

“You have the Fed embarking on a long, slow tightening cycle. Tightening into a profits recession doesn’t sound like anything to throw a big party about,” she said.

Ben Carlson saw her comments and decided to take a closer look at the historical relationship between profit growth and stock market returns.

One of the biggest problems in the world of finance is that people make proclamations without backing it up with evidence. So I wanted to see what the historical relationship looks like between profit growth and stock market returns. Using Federal Reserve data on corporate profits, I looked back at the historical growth rate of profits by decade and compared them to that decade’s stock market returns (using the S&P 500):

Carlson1

Now let’s break things down even further by market type:

Carlson2

(Although the S&P 500 was up 6% per year in the 1966-1981 period, many consider this a sideways market because the Dow went nowhere from a price perspective and once you take inflation into account real returns were basically zero.)

There’s really not much of a discernible pattern that can be detected here. High profit growth has led to both high and low stock market returns throughout the post-WWII period. There were also times of low profit growth with high stock market returns.

The greatest profit growth was seen in two of the worst-performing stock market decades — the 1970s and 2000s. But those periods were markedly different as the 70s saw sky-high inflation with rising interest rates while the 2000s had low inflation and falling rates.

After accounting for inflation, the 1980s only saw profit growth of roughly 1.6%, but stocks returned more than 17% per year (12% real). The 1950s and 1960s saw one of the greatest bull markets of all-time, but profit growth was basically average. Profits growth has been non-existent during the latest bull market cycle, but stocks are up gangbusters anyways.

This may seem highly blasphemous to die hard fundamentalists who often fall into the trap of thinking that success in the stock market is mostly a matter of accounting.  It is not.  Rather, it all comes back to what investors are willing to pay for those earnings.  Thus the need for technical analysis!

Posted by:


Sector Performance

March 24, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 3/23/16.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Sector Performance

March 11, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 3/10/16.

sector

The performance above is based on pure price returns, not inclusive of dividends, fees, or other expenses.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.  Source: iShares

Posted by:


Sector Performance

February 26, 2016

The table below shows performance of US sectors over the trailing 12, 6, and 1 month(s).  Performance updated through 2/25/15.

sector

The performance above is based on pure price returns, not inclusive of dividends or all transaction costs.  Past performance is not indicative of future results.  Potential for profits is accompanied by possibility of loss.    Source: iShares

Posted by: