A Homemade 130/30 Investment Strategy: A Longer-Term Perspective [View article]
Thanks for reading!
Brian: the current beta on TLT is much closer to -.78 as you suggest - which is what I show in my last article about 130/30 (linked at the top of this article). The only reason it is not nearly as negative in this analysis is that this analysis covers a much larger time frame.
This article, vs. the prior article, doesn't provide quite as much clarity about my objective - which was to provide an alternative strategy to achieve the desired result of constructing a 130/30 portfolio without actually short selling any securities.
As I think about it, unless you are exceptionally talented at stock picking, the purpose of short selling in a 130/30 portfolio is to constrain net market exposure to roughly 100% while attempting to generate some excess return via security selection. The purpose of holding TLT, LQD, and GLD in this context, is that they exhibit low and even negative Betas which effectively reduces the portfolios net market exposure (measured by Beta) to 100% while offering the potential to earn excess returns on security selection via duration, interest rate, and inflation risk exposures. I was not really thinking about them as a means to lever up the portfolio to 130% - after all the purpose of 130/30 is to maintain net market exposure of 100%.
Your market view on TLT is definitely worth considering. Although, I think it is challenging to make a case that interest rates can rise dramatically without a material improvement in domestic/global economic conditions - which, theoretically, should result in higher stock prices despite increased inflation.
GLD Investor: the weight adjustment process was annual after the first 14 months. Since the data was available starting in November, I waited until the end of the next calendar year to conduct the first adjustment - after which it was an annual event at the end of each calendar year.
The process was to tweak the holding weights of QQQ & TLT (and to a lesser extent GLD & LQD) such that over the previous 12 months, using the adjusted holding weights, the portfolio would have exhibited a beta of 1.0 plus or minus roughly 5%. The adjusted holding weights were used over the next 12 month time period and this was repeated 8 times. Realistically my adjustment process was a fairly naive methodology. In real life you will want to incorporate market views on each security whereas I very blindly focused on retroactively hitting the beta number and carrying it forward to simply see the results.
Hopefully, this answers your question sufficiently, if not hit me up again and I will try to clarify further.
A Homemade 130/30 Investment Strategy: A Longer-Term Perspective [View article]
My apologies if it appears I am implying that this strategy will outperform the S&P 500 over the next 40 years. That is not at all what I meant to do. My intent was to numerically re-iterate the importance of compounding and how a seemingly small positive differential can result in a dramatically better outcome (provided it is sustainable over a specific period) - not to suggest that THIS strategy WILL perform that well over the next 40 years.
Asset Inflation: Correlations Are All Positive [View article]
Mmm...Perhaps you are measuring correlation in a unique way...but when I calculate it, the correlation I am coming up with is -.64 on TLT & SPY using daily returns from June 2012 through December 10 for my data set. Additionally, using a 30 day rolling correlation of TLT & SPY from June 2011 through December 12, the maximum 30 day correlation per my calculation was -.13 (which is still negatively correlated).
Volatility Is A Risk Worth Shorting [View article]
An excellent build/extension of the discussion. This provides a nice teaser on Risk Parity - which, based on several of the comments that were posted on my article, as well as Roger Nusbaum's article, remains a methodology that appears to be somewhat widely misunderstood and confusing. Good job - we have to keep it going!
Is Everything We Know About Stock/Bond Allocations Wrong? [View article]
I accept.
P.S. I am flattered that you found my article interesting/influential enough that you 1.) read it more than once and 2.) found it compelling enough to invest the time required to write an entire blog post devoted to raising questions and posing critiques to several of the arguments in my article.
Will The Real 60/40 Please Stand Up? [View article]
Haha. That is good. I wonder if someone were to write about the impact of divorce on retirement outcomes if SA would publish it.
Hayekvonfriedman - that is actually a complicated set of questions and a thorough/complete answer will require a substantial amount of work. My process actually assumes monthly re-balancing for better or worse. I don't know if this is the best re-balancing period or even if changing re-balancing periods has a material impact on this analysis. Intuitively, it may impact the performance data shown at the bottom but that historical performance is just that anyways: historical. On a high level, optimal re-balancing is essentially analogous to picking winners - just on a much smaller scale. The important thing is that you are maintaining the approximate risk exposures you desire overtime and the specifics of how large a deviation it takes to warrant a re-balancing is just one opinion vs. another. I am sure there are tons of research papers that explore optimal re-balancing and what impact it has on various strategies.
With respect to transaction costs, I totally ignore them - which I would argue they are immaterial anyways for a strategy like this given the ultra low cost of ETF transaction fees at various low cost brokers.
Your questions regarding secular or cyclical trends are relevant but unfortunately equivalent to asking which asset classes will outperform on a risk adjusted basis in the future...e.g. how do demographic factors impact financial flows into various asset classes (i.e. equity valuations, performance, volatility, etc.)? That question has tons of merit, but it is a market call and in effect the outcome would likely involve skewing a portfolio's risk allocation marginally in favor of which ever asset class you expect to perform best (on a risk adjusted basis) due to the secular trends vs. other asset classes. I am sure there are countless research papers evaluating the role of strategic asset allocation (long term secular trends) vs. tactical asset allocation (cyclical trends) and maybe they could be of some value.
The most interesting part of your comment, in my opinion, is the question regarding contingency plans for the extremely low interest rate environment and how to best re-allocate as interest rates normalize to more historic levels. That is an awesome question and the truth is I don't know. My initial thought is that when the facts change, you re-balance which is to say when the "fixed weights" no longer provide the desired risk balance between equities and fixed income, you adjust the weights. The more important issue that I don't know how to resolve is this: as interest rates rise the diversifying properties of bonds will increase. This is due to the fact that at the point long term interest rates are higher vs. current levels there will be room for them to return to low levels again in the event of a market downturn - effectively leading to price appreciation and potentially strong relative out-performance of bond funds vs. equities. Additionally, there are other fixed income funds such as 1-3 year, 4-7 year duration funds that may become the superior fixed income instruments, from a statistical characteristics perspective, to include in the asset mix during a period of rising interest rates. This may actually turnout to be the best answer...as interest rates begin to actually rise, an investor should re-allocate more heavily to the front end of the yield curve and take on more credit risk as opposed to duration and interest rate risk for their fixed income allocation.
There are probably people on here much more capable than me to actually explain the mathematics, as well as the relative importance of risk factors such as credit, duration, interest rate etc., and demonstrate how these varying factors can play out in varying scenarios and perhaps provide some good guidance for the future.
Studies of risk parity strategies (which is what my analysis is effectively based on) have shown robust performance over long investment horizons. It's during periods of significant one way equity out-performance vs. other assets that this approach to asset allocation looks less than desirable - but that becomes a question of timing capabilities.
Will The Real 60/40 Please Stand Up? [View article]
68osb28 - congrats on the good performance doing it yourself. Having said that just remember that any asset allocation mix that incorporates multiple asset classes will likely under perform a portfolio 100% allocated to the strongest performing asset class during a set time period using only return driven metrics - it's when that cycle turns that you get crushed. If you happen to be continuously blessed with fortuitous timing skills and can successfully re-allocate your asset mix prior to a severe correction than you will no doubt fare very well.
I have noticed TONS of investment firms, mutual funds, etc use very bastardized versions of MPT and from what I can tell, a large portion of them are essentially mediocre versions of the S&P 500 - literally by design and perverse incentive structures.
Will The Real 60/40 Please Stand Up? [View article]
Old Trader - My absolute two favorite investing books are Steven Drobny's "Inside the House of Money" and "The Invisible Hands". At 27 years old, they have been immensely influential on my way of thinking. There are so many thought provoking points made by the fund managers interviewed in his books. I can't tell you how many times I have re-read chapters only to gain new perspectives and potential ideas that I never considered in my 2nd or 3rd reading of those same chapters. And yes, the concepts underlying my article are pretty much exactly what "the pensioner" discussed in that book.
Will The Real 60/40 Please Stand Up? [View article]
Smarty_Pants: You make a valid point. The time period under consideration is limited and potentially skewed in such a way that the performance chart at the bottom of the article is not necessarily indicative of what the next 5 years will look like. That does not in any way invalidate the approach or conceptual premises underlying this article. It simply introduces additional information that needs to be carefully thought through when making any investment decisions.
Toledoinvestor: my time frame was constrained by data availability on the specific funds I incorporated into the analysis.
You raise some important issues in your first comment. When I discuss risk in this article I am referring to volatility. And when I discuss risk exposure I am referring to the overall portfolio volatility that can be attributed to each asset class (in retrospect, I should have called this "risk contribution" to help prevent any confusion). I don't explicitly discuss risk factors (i.e. credit, duration, liquidity, etc).
With respect to the relative correlation of each fund vs. the S&P 500, that all depends on the time frame you are using to measure the correlation. Over the entire time period in question I am showing a correlation to SPY of -.10 for AGG and 0.168 for LQD (using daily returns from Jan 1 - Sept 6). The higher level point that I would argue you are raising is that over the time period in question, AGG has not exhibited enough volatility within its available risk factors (credit, duration, liquidity, etc.) to meaningfully contribute to the portfolio's volatility vs. the volatility exhibited by stocks over the same time frame. This is exactly why the risk exposure calculation using only SPY and AGG yields a risk exposure of 95% to equities and 5% to fixed income.
Moreover, only after increasing the various risk factors you highlight i.e. credit, duration, etc., did the fixed income allocation exhibit sufficient volatility to actually contribute to the volatility of the overall portfolio - which is to say it actually reflected a true 60/40 asset allocation.
And with respect to showing my calculations: only because I don't actually run a fund : )
I want to provide the calculation, but it is pretty difficult to type it in the comment box without looking pretty confusing. So in lieu of an typed out formula, I just found this slide show put out by Panagora that explains the calculation. Incidentally if you look on slide 6 (pg 7) they show the same thing I have shown with the 95% and 5% risk exposure/contribution breakdown between stocks and bonds. They go into a lot more stuff in this slide show but hopefully this will help you out as far as calculations go. Let me know if this is insufficient for providing the calculation I used, and I will figure out how to type it out without confusing everybody involved.
Will The Real 60/40 Please Stand Up? [View article]
Good comments from everyone. Let me start from the bottom and move towards the top:
1.) Mgordon10 - On what basis are you arguing that the classic 60/40 portfolio is far superior to the alternatives I purpose? I would argue fixed income investments should provide far more than a simple anchor to a risky (read: volatile) equity portfolio. The implication of your argument, if I have understood it correctly, is that investing is 100% about stock picking - as fixed income is included to simply work as an anchor against the excessive volatility of stocks vs. including fixed income as a means of introducing incremental returns/diversification properties to the overall portfolio. Which, if I have understood your argument correctly, I can't agree. Depending on the mathematics, you can pick a majority of "winning" stocks and still end up worse off than had you invested in a diversified portfolio of different asset classes.
2.) Geneh - I am not entirely sure I understand your question. If you wanted to go over the top and really do this the right way you would include every asset and associated cash flow/return stream ranging from SS, equities, real-estate, etc. Having said that, I did not specifically include SS or annuities into the "fixed income" allocations.
3.) Crankyguy - Let me ask you this...how do your investors react to modestly under performing the market during up swings? I have full confidence that this strategy works over a long investment horizon but during periods when equities go up 15-25%, this strategy would likely under-perform an all-equity portfolio and I just wonder how quick a client is to fire you as they tend to focus only on upside that they feel they are missing out on effectively forgetting what happens during bear markets?
4.) Jonathan - I read the Bridgewater whitepaper a while back...and that paper as well as a few others completely changed my life. In fact, I have incorporated risk parity concepts into many of my SA contributions.
5.) I will make a specific comment for Toledoinvestor.
6.) Varan - I don't know anything about the Morning Star fund selector tool but I do agree that many people got crushed because they did not have a thorough enough understanding of the risks inherent in their portfolios.
Will The Real 60/40 Please Stand Up? [View article]
I appreciate all of the comments and praise. Let me answer the question regarding relevancy with another question: does diversification matter?
If diversification does not matter, then this is not relevant to anyone regardless of their age. If diversification does matter, then the concepts underlying this discussion are not only relevant, they are somewhat crucial to success.
While I chose to focus on asset allocation between fixed income and equities in order to extend on another SA contributor's article, this same analysis can readily be applied to any asset mix under review-including 100% equity portfolios.
Will The Real 60/40 Please Stand Up? [View article]
A Homemade 130/30 Investment Strategy: A Longer-Term Perspective [View article]
Brian: the current beta on TLT is much closer to -.78 as you suggest - which is what I show in my last article about 130/30 (linked at the top of this article). The only reason it is not nearly as negative in this analysis is that this analysis covers a much larger time frame.
This article, vs. the prior article, doesn't provide quite as much clarity about my objective - which was to provide an alternative strategy to achieve the desired result of constructing a 130/30 portfolio without actually short selling any securities.
As I think about it, unless you are exceptionally talented at stock picking, the purpose of short selling in a 130/30 portfolio is to constrain net market exposure to roughly 100% while attempting to generate some excess return via security selection. The purpose of holding TLT, LQD, and GLD in this context, is that they exhibit low and even negative Betas which effectively reduces the portfolios net market exposure (measured by Beta) to 100% while offering the potential to earn excess returns on security selection via duration, interest rate, and inflation risk exposures. I was not really thinking about them as a means to lever up the portfolio to 130% - after all the purpose of 130/30 is to maintain net market exposure of 100%.
Your market view on TLT is definitely worth considering. Although, I think it is challenging to make a case that interest rates can rise dramatically without a material improvement in domestic/global economic conditions - which, theoretically, should result in higher stock prices despite increased inflation.
GLD Investor: the weight adjustment process was annual after the first 14 months. Since the data was available starting in November, I waited until the end of the next calendar year to conduct the first adjustment - after which it was an annual event at the end of each calendar year.
The process was to tweak the holding weights of QQQ & TLT (and to a lesser extent GLD & LQD) such that over the previous 12 months, using the adjusted holding weights, the portfolio would have exhibited a beta of 1.0 plus or minus roughly 5%. The adjusted holding weights were used over the next 12 month time period and this was repeated 8 times. Realistically my adjustment process was a fairly naive methodology. In real life you will want to incorporate market views on each security whereas I very blindly focused on retroactively hitting the beta number and carrying it forward to simply see the results.
Hopefully, this answers your question sufficiently, if not hit me up again and I will try to clarify further.
A Homemade 130/30 Investment Strategy: A Longer-Term Perspective [View article]
Have You Ever Considered 130/30? [View article]
Asset Inflation: Correlations Are All Positive [View article]
What am I missing?
Is Low Vol A Beta Phenomenon? [View article]
Was that a typo? Did you mean to put it is indisputable that low beta OUTperforms high beta?
Volatility Is A Risk Worth Shorting [View article]
Is Everything We Know About Stock/Bond Allocations Wrong? [View article]
Is Everything We Know About Stock/Bond Allocations Wrong? [View article]
P.S. I am flattered that you found my article interesting/influential enough that you 1.) read it more than once and 2.) found it compelling enough to invest the time required to write an entire blog post devoted to raising questions and posing critiques to several of the arguments in my article.
Will The Real 60/40 Please Stand Up? [View article]
Hayekvonfriedman - that is actually a complicated set of questions and a thorough/complete answer will require a substantial amount of work. My process actually assumes monthly re-balancing for better or worse. I don't know if this is the best re-balancing period or even if changing re-balancing periods has a material impact on this analysis. Intuitively, it may impact the performance data shown at the bottom but that historical performance is just that anyways: historical. On a high level, optimal re-balancing is essentially analogous to picking winners - just on a much smaller scale. The important thing is that you are maintaining the approximate risk exposures you desire overtime and the specifics of how large a deviation it takes to warrant a re-balancing is just one opinion vs. another. I am sure there are tons of research papers that explore optimal re-balancing and what impact it has on various strategies.
With respect to transaction costs, I totally ignore them - which I would argue they are immaterial anyways for a strategy like this given the ultra low cost of ETF transaction fees at various low cost brokers.
Your questions regarding secular or cyclical trends are relevant but unfortunately equivalent to asking which asset classes will outperform on a risk adjusted basis in the future...e.g. how do demographic factors impact financial flows into various asset classes (i.e. equity valuations, performance, volatility, etc.)? That question has tons of merit, but it is a market call and in effect the outcome would likely involve skewing a portfolio's risk allocation marginally in favor of which ever asset class you expect to perform best (on a risk adjusted basis) due to the secular trends vs. other asset classes. I am sure there are countless research papers evaluating the role of strategic asset allocation (long term secular trends) vs. tactical asset allocation (cyclical trends) and maybe they could be of some value.
The most interesting part of your comment, in my opinion, is the question regarding contingency plans for the extremely low interest rate environment and how to best re-allocate as interest rates normalize to more historic levels. That is an awesome question and the truth is I don't know. My initial thought is that when the facts change, you re-balance which is to say when the "fixed weights" no longer provide the desired risk balance between equities and fixed income, you adjust the weights. The more important issue that I don't know how to resolve is this: as interest rates rise the diversifying properties of bonds will increase. This is due to the fact that at the point long term interest rates are higher vs. current levels there will be room for them to return to low levels again in the event of a market downturn - effectively leading to price appreciation and potentially strong relative out-performance of bond funds vs. equities. Additionally, there are other fixed income funds such as 1-3 year, 4-7 year duration funds that may become the superior fixed income instruments, from a statistical characteristics perspective, to include in the asset mix during a period of rising interest rates. This may actually turnout to be the best answer...as interest rates begin to actually rise, an investor should re-allocate more heavily to the front end of the yield curve and take on more credit risk as opposed to duration and interest rate risk for their fixed income allocation.
There are probably people on here much more capable than me to actually explain the mathematics, as well as the relative importance of risk factors such as credit, duration, interest rate etc., and demonstrate how these varying factors can play out in varying scenarios and perhaps provide some good guidance for the future.
Studies of risk parity strategies (which is what my analysis is effectively based on) have shown robust performance over long investment horizons. It's during periods of significant one way equity out-performance vs. other assets that this approach to asset allocation looks less than desirable - but that becomes a question of timing capabilities.
Will The Real 60/40 Please Stand Up? [View article]
I have noticed TONS of investment firms, mutual funds, etc use very bastardized versions of MPT and from what I can tell, a large portion of them are essentially mediocre versions of the S&P 500 - literally by design and perverse incentive structures.
Will The Real 60/40 Please Stand Up? [View article]
Will The Real 60/40 Please Stand Up? [View article]
Toledoinvestor: my time frame was constrained by data availability on the specific funds I incorporated into the analysis.
You raise some important issues in your first comment. When I discuss risk in this article I am referring to volatility. And when I discuss risk exposure I am referring to the overall portfolio volatility that can be attributed to each asset class (in retrospect, I should have called this "risk contribution" to help prevent any confusion). I don't explicitly discuss risk factors (i.e. credit, duration, liquidity, etc).
With respect to the relative correlation of each fund vs. the S&P 500, that all depends on the time frame you are using to measure the correlation. Over the entire time period in question I am showing a correlation to SPY of -.10 for AGG and 0.168 for LQD (using daily returns from Jan 1 - Sept 6). The higher level point that I would argue you are raising is that over the time period in question, AGG has not exhibited enough volatility within its available risk factors (credit, duration, liquidity, etc.) to meaningfully contribute to the portfolio's volatility vs. the volatility exhibited by stocks over the same time frame. This is exactly why the risk exposure calculation using only SPY and AGG yields a risk exposure of 95% to equities and 5% to fixed income.
Moreover, only after increasing the various risk factors you highlight i.e. credit, duration, etc., did the fixed income allocation exhibit sufficient volatility to actually contribute to the volatility of the overall portfolio - which is to say it actually reflected a true 60/40 asset allocation.
And with respect to showing my calculations: only because I don't actually run a fund : )
I want to provide the calculation, but it is pretty difficult to type it in the comment box without looking pretty confusing. So in lieu of an typed out formula, I just found this slide show put out by Panagora that explains the calculation. Incidentally if you look on slide 6 (pg 7) they show the same thing I have shown with the 95% and 5% risk exposure/contribution breakdown between stocks and bonds. They go into a lot more stuff in this slide show but hopefully this will help you out as far as calculations go. Let me know if this is insufficient for providing the calculation I used, and I will figure out how to type it out without confusing everybody involved.
http://bit.ly/Rt9nG1
Thanks again for reading and all the comments. Good discussion.
Will The Real 60/40 Please Stand Up? [View article]
1.) Mgordon10 - On what basis are you arguing that the classic 60/40 portfolio is far superior to the alternatives I purpose? I would argue fixed income investments should provide far more than a simple anchor to a risky (read: volatile) equity portfolio. The implication of your argument, if I have understood it correctly, is that investing is 100% about stock picking - as fixed income is included to simply work as an anchor against the excessive volatility of stocks vs. including fixed income as a means of introducing incremental returns/diversification properties to the overall portfolio. Which, if I have understood your argument correctly, I can't agree. Depending on the mathematics, you can pick a majority of "winning" stocks and still end up worse off than had you invested in a diversified portfolio of different asset classes.
2.) Geneh - I am not entirely sure I understand your question. If you wanted to go over the top and really do this the right way you would include every asset and associated cash flow/return stream ranging from SS, equities, real-estate, etc. Having said that, I did not specifically include SS or annuities into the "fixed income" allocations.
3.) Crankyguy - Let me ask you this...how do your investors react to modestly under performing the market during up swings? I have full confidence that this strategy works over a long investment horizon but during periods when equities go up 15-25%, this strategy would likely under-perform an all-equity portfolio and I just wonder how quick a client is to fire you as they tend to focus only on upside that they feel they are missing out on effectively forgetting what happens during bear markets?
4.) Jonathan - I read the Bridgewater whitepaper a while back...and that paper as well as a few others completely changed my life. In fact, I have incorporated risk parity concepts into many of my SA contributions.
5.) I will make a specific comment for Toledoinvestor.
6.) Varan - I don't know anything about the Morning Star fund selector tool but I do agree that many people got crushed because they did not have a thorough enough understanding of the risks inherent in their portfolios.
Thanks again for reading and all of the comments.
Will The Real 60/40 Please Stand Up? [View article]
If diversification does not matter, then this is not relevant to anyone regardless of their age. If diversification does matter, then the concepts underlying this discussion are not only relevant, they are somewhat crucial to success.
While I chose to focus on asset allocation between fixed income and equities in order to extend on another SA contributor's article, this same analysis can readily be applied to any asset mix under review-including 100% equity portfolios.
Thanks again for reading.