Cleaning Out My Portfolio: In Defense Of The Real 60/40

|
 |  Includes: GLD, HYG, MBB, SPY
by: Daniel Moser

Last weekend, I wrote an article called "Will the Real 60/40 Please Stand Up?" which Seeking Alpha was kind enough to publish on September 8th. My inspiration for this article came from an article written by Matt Tucker which I had read earlier that week. In Mr. Tucker's article, he essentially started a conversation about what role fixed income could play in an investor's portfolio, and he suggested that investors-young and old alike-should consider something along the lines of the classic 60/40 asset allocation (60% allocated to equities, 40% allocated to fixed income). I extended on that discussion by arguing that the classic 60/40 portfolio does not actually represent a true 60/40 asset allocation if you consider the portfolio from a "risk" perspective.

Since my article was published, a flurry of readers posted some great comments, raised important questions, and posited counter arguments to several of the underlying ideas within my article that are no doubt worth consideration. One reader and fellow Seeking Alpha contributor, Roger Nusbaum, graciously took the time to read my article and was compelled enough to invest his time in writing an entire Seeking Alpha article devoted to raising questions about my article, pointing out areas that he felt I did not adequately explain, and highlighting areas of disagreement that readers/investors should consider. I would like to thank him for investing his time to extending this discussion further and, naturally, I feel compelled to follow up on his critique and refute several of his points of contention. So, let the games begin.

Mr. Nusbaum begins his foray into my article by raising a concern surrounding some ambiguity that I definitely think warrants some clarification, based on comments and his critique, in order to have an effective discussion: what do I consider risk? In my article, I wrote:

For the purpose of this article 'risk exposure' is the percentage of portfolio volatility that is attributable to each asset class.

First of all, in retrospect, I should have called this "risk contribution" rather than "risk exposure." Either way, let me clarify that point right now. I am arguing that risk is equal to volatility. I know that several highly intelligent people, with terrific investment track records, take issue with this point of view. Having said that, I am not really interested in having a debate with a central focus on the merits of volatility as risk vs. probability of loss as risk or whatever other definition of risk one wishes to consider. Thus, I will simply posit that the single largest risk an investor faces, from a broad perspective, is that he or she fails to maximize wealth accumulation by inadequately compounding returns over their respective investment time horizon - which is to say, anything that destroys compounding poses risk to the investor. It is a mathematically provable fact that volatility erodes the impacts of compound interest.

Moreover, defining the underlying risk factors inherent in an investment opportunity (credit, duration, equity, currency, etc.) is woefully inadequate with respect to portfolio construction, largely because an investor still faces the question of how much of their portfolio should be allocated to those defined risk factors. As such, I would simply argue that volatility serves as a pretty good tool, worth considering, in determining the appropriate level of allocation to each investment/asset class in order to achieve an optimally diversified portfolio.

Next, Mr. Nusbaum stages his second point of criticism by arguing that I fail to address "why it makes for poor portfolio construction to have 95% of the risk isolated to equity exposure." My first reaction to this was: I told you that tons of professional investors consider a dollar cost percentage weighted 60/40 asset allocation adequately diversified, and he appears to fit into this camp. Nonetheless, he makes a valid point. I did not explicitly state why having 95% of a portfolio's risk isolated to equities is a bad thing. Quite simply, I am arguing that a 95% risk allocation to equities is not an optimally diversified portfolio. For a moment, let's consider Mr. Nusbaum's counter point that he argued later on in his article: that portfolio construction is difficult because if you go too aggressive the "risk becomes panic selling at a generational low but going too conservative could result, of course, in coming up short." I beg to differ. Who cares if you liquidate at a generational low - I am far more concerned about losing that much wealth in the first place! Okay, I have to be somewhat reasonable by conceding that liquidating a portfolio at a generational low is probably a bad thing. The unfortunate fact for Mr. Nusbaum is that he places himself in a quasi-paradox by totally neglecting the impact of compounding returns. Returns are not symmetrical. A portfolio losing 50% of its value in one year must subsequently experience investment returns of 100% in order to return to its previous value. Quite simply, in order to maximize compounding, an investor must avoid large draw-downs. Large draw-downs occur by investing in a portfolio with a large allocation to highly correlated and risky (e.g. volatile) assets that experience a significant correction simultaneously a.k.a. sub-optimal or even inadequate diversification.

While my example may seem a little extreme, consider for a moment that the S&P 500 fell nearly 60% from the highs of 2007 to the lows in 2009. Let's just say (and please knock on wood) that history repeats itself and equities take a plunge of 60%. Afterwards, the market subsequently realizes returns something a little bit closer to the historical average (call it 12% per year - generous but certainly more typical than 25% per year). It would take roughly 8 years to recoup losses of that sort of magnitude. Now, in fairness, additional contributions of capital throughout the full investment cycle can dramatically reduce the time required to recoup losses.

Mr. Nusbaum extends his critique further, in a third wave, by arguing that my theory does not hold water because:

1.) "Depending on how [much] more risk is taken in the fixed income market you might as well be in equities"
2.) "Many people own fixed income to offset normal stock market volatility."

Again, I must respectfully disagree…we just don't see this the same way.

With respect to his first point on his third assault, while it is true that particular risk factors inherent in various fixed income instruments do exhibit a positive correlation to equities (i.e. high yield bond credit risk vs. equities, etc.), these risk factors are rarely perfectly correlated - which is to say there are diversification benefits available to investors who allocate capital in some combination of fixed income instruments, such as high yield debt and equities vs. a 100% allocation to equities. In fact, based on daily returns from Jan 2008 through year to date 2012, the maximum 90 day correlation that the high yield bond ETF (NYSEARCA:HYG) exhibited to the S&P 500 (NYSEARCA:SPY) was 0.885. While this is certainly high, it is not 1.00, which means that even at the peak of correlation, an investor could still reduce their portfolio's volatility, thereby improving risk adjusted returns, by diversifying into HYG. As it turns out, the minimum level of 90-day correlation over this same time period was 0.175, which, surely everyone will agree, implies that a significant amount of potential diversification benefits are available to the investor who allocates some portion of capital to HYG and the rest to SPY.

In refutation to his second point, on his third wave (and I did find myself paraphrasing the exact wording of his argument to what I thought he was trying to argue), despite the fact that some fixed income instruments behave more like equities in comparison to other fixed income instruments, 1.) They don't behave exactly like equities and moreover 2.) The fixed income instruments that don't tend to exhibit a positive relationship to equities (i.e. long-term government debt) will actually mitigate poor equity returns through their empirically provable inverse-return relationship. By increasing a portfolio's risk exposure (risk contribution) to these latter instruments, an investor will actually tend to dramatically reduce the overall portfolio volatility while simultaneously increasing the level of diversification as the risk factors inherent in the overall portfolio will tend to be unique from one another and in some instances will exhibit offsetting statistical behaviors to the risk factors inherent in equities. Recall, volatility destroys compounding, thus reducing volatility is actually equivalent (to a point) of increasing returns over time as lower volatility supports maximum compounding.

Mr. Nusbaum's fourth layer of arguments begins a drift into a discussion that goes beyond the scope of my original article (but I welcome the conversation anyways), and in doing so Mr. Nusbaum poses an argument that in my mind is contradictory to his overall point of view. He states towards the beginning of his article that I appear to be "doing something that Cliff Asness from AQR has talked about, which is allocating risk though not necessarily allocating asset classes." Mr. Nusbaum argues that I just have it wrong by stating that, "I would say that I do not believe that the risk needs to be allocated in the manner that I think Moser is talking about." But, he goes on to say that, "if the conversation gravitates to several different asset classes, as Asness discusses, then that could very well be a different story." This poses an inherent contradiction due to the fact that, in my mind, a risk based asset allocation process either works or it does not. Why are readers to believe that a risk based portfolio construction process may in fact be a sound methodology for portfolio construction when discussed in the context of 5 different asset classes, but somehow it is an unsound methodology in the context of 2 asset classes? Is diversification really less important in the context of 2 asset classes vs. 3, 4, or even 5 asset classes? I can't possibly believe this to be true…and neither should the readers. Furthermore, I must respectfully suggest that the reason Mr. Nusbaum stands ready to dismiss my proposed risk-based methodology of asset allocation in the context of 2 asset classes, but remains open to accepting it in the context of 5 asset classes, is that he may have failed to adequately understand a risk-based portfolio allocation methodology. The underlying supposition in my article is that a dollar percentage weighted portfolio leads to sub-optimal diversification which at best reduces compounding over a long investment time horizon and at worst exposes investors to an increased likelihood of a major compound killing draw-down in the event of a material equity market downturn - even when considered within the context of 2 asset classes.

And just to be clear, in my article, I only consider fixed income (in a semi-generic sense) and equities for asset classes to incorporate into my article because my intent was to demonstrate that the classic 60/40 asset allocation does not actually create a diversified asset allocation mix from a risk perspective. I could have incorporated additional asset classes such as gold (NYSEARCA:GLD) or other slices of the fixed income pie such as mortgage bonds (NYSEARCA:MBB), etc. which tend to exhibit a very weak relationship to equities, thus, generating even more opportunities to enhance the diversification of a portfolio. With all of that said, the specific assets I included in my article and calculations are not really the central focus of my article…the process is.

In his fifth foray into my article, Mr. Nusbaum asserts that I failed to consider that risk changes over time. However, towards the bottom of my article, I did write that, "just in case it does not go without saying, there is certainly no guarantee that the exact weights of the portfolio I refer to as the true 60/40 portfolio will actually behave like a true 60/40 portfolio in the future - which is why a portfolio must be re-balanced overtime." There is really nothing more to say here. I am confident he was simply more focused on other areas of my article that this was unintentionally overlooked. Just in case this was not an unintentional error, I unequivocally believe that re-balancing a portfolio over time is crucial to the success of virtually every investment strategy, as the facts and data are in constant flux.

At this point, I think I have refuted virtually every argument Mr. Nusbaum raised concerning my article. The cold, harsh reality is that no one really knows when the next big crises, meltdown, or correction will occur. People speculate, ponder, and muse, but at the end of the day it all amounts to a guess. It may indeed be a well thought out guess based on terrific logic and countless facts or it may be a complete hunch - either way, it amounts to little more than a guess. In my opinion, constructing a portfolio utilizing a risk focused process is far superior to a dollar cost percentage weighted asset allocation methodology, largely because I can be honest/modest enough with myself to suggest that, while I may be willing to hazard a guess from time to time, I really don't know what is going to happen next. Thus, why would I ever prefer a portfolio that concentrates the VAST majority of its inherent risk among relatively few unique assets? I suspect that practically every reader on Seeking Alpha has heard the phrase "don't put all your eggs in one basket." I am left wondering if it is time for investors to take a step back and ask a basic question: what is an egg?

Thanks for reading (especially if you made it to the end!) and good luck.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.