Modern Portfolio Theory turned upside down

Let's start with a brief discussion of modern portfolio theory (MPT)  just to make sure we're all on the same page.   To avoid a less than exciting academic foray into MPT, I'll provide a basic definition.

Basically, the theory is that individual investment (a particular asset i.e. stock or bond) selection should not be chosen on their own merits. What's more important is how each asset's price moves relative to all the other investments in a portfolio.  Harry Markowitz was the first to articulate this concept back in 1952.

The market is one of extremes and MPT is seen as the answer to those extremes. Experts will tell you all day long that the average annual return in the market for the last 100 years is between 8-10% (depending on what they are calling “the market”).  But in reality, there are very few years where the market returns 8% as an example.

Care to guess how many times the average annual return has been between 8-12%?

Well, since 1926 it's only happened five times.

Yep, that's right.  In the last 86 years, the stock market has had five years that were average.

So, what's happened the the other 81 times?  It's a mixed bag of extreme highs and heart-stopping lows.

stock market rollercoaster

Okay now to get more specific, let’s look at a chart that I like to call “the periodic table of investment returns”.

This one is constructed using data from Lipper (a third party company that compiles market data) and it gives us a snapshot of how a “diversified portfolio” performed for the last 10 years (2002-2011).

modern portfolio theory


These charts are put together to prove to you, the consumer, how undeniably awesome modern portfolio theory works (MPT) and how the use of a balanced or diversified portfolio is the greatest idea Wall Street has every conceived.

In typical insurance pro blog style, we’re going to use all the wonderful research that Wall St. has commissioned and use it to tell a different story.

Does modern portfolio theory really work?

Let’s do some math.

Certainly, we can all see that over this particular 10 year period we have varying rates of return from year-to-year as one would expect.  If we calculate the average annual return (arithmetic mean) we get 7.3%, however, if you go read my post on compound annual growth rate (CAGR) then you’ll understand that’s not the number we should be concerned with.

The CAGR is only 5.6% which is the number that really matters to anyone who actually has money invested.  Only the Wall St. marketing machine cares to mention average annual return numbers.

Why is that?

Because they’re allowed to use those numbers in promotional literature and they always look better than CAGR.

So, this begs the question.  Why would I risk having a -31% year like we had in 2008 to get a 5.6% compound annual growth rate over ten years, when I could have gotten a similar return  inside of a good participating whole life contract or indexed universal life contract?

Wouldn’t it be better to take advantage of the stable values inside of an insurance contract and have the ability to purchase stocks when market is tanking?  After all, the financial experts are always telling us that the perfect time to invest in the market is when the market takes a significant drop.

Remember what Warren Buffet said:

“Be fearful when others are greedy, be greedy when others are fearful”

I believe this should be the basis for all investment decision, however, it's easier said than done.  I don't know about you, but I'm pretty emotional about my money?

And I hate losing it.  That's why I don't gamble.  Not passing judgment on those who enjoy Vegas or poker night with friends but it feels as natural to me as dumping boxes of cash in my fireplace to start a fire.

And many financial planner types will tell you, “A well-diversified portfolio is your best defense against market volatility.”


You wanna know the best defense against market volatility?

Decrease your exposure to the market.  Pretty simple.  The less money you have in the market, the less chance you'll have of experiencing the wild gyrations that come along with it.

If I’m continually investing into a strategy that has the potential to fall precipitously at any moment without any warning, then I don’t have much opportunity to buy into the dips.

In my past life, I saw the impact of a retired person beginning to take withdrawals from a diversified portfolio.  Traditional financial planning says that when you’re going to retire and begin taking income from your portfolio, you should pick a coupon rate…say 4-5% that you will withdraw from your investments every year, no matter what your return is for any given year.

monte carlo simulationSidenote: There’s even fancy software that will run a bazillion scenarios, called monte carlo simulations to project the probability of you running out of money based on historical returns of a particular portfolio mix.  I don’t want to wander too far down that dark and winding road because mostly it’s purely academic and has no basis in reality.

What no one ever thinks about is…

What happens if I started my retirement in 2008 and the first year I had been down 30%?

I’ll tell you what happens.

It’s a complete disaster for your retirement if you’re using the method of income as I mentioned above.  You need must have stable values that don’t fluctuate wildly during your non-income producing years.  Unfortunately, no one plot the expansions and contractions of the market so that you can benchmark it to your retirement plan, that’s just not reality based planning.

So, if the compound annual growth rates are basically the same in a high quality participating whole life insurance policy or indexed universal life insurance policy, why bother risking the crazy ride?  It’s better just to have the stability in your plan.

If you’d like to learn more about our view of modern portfolio theory or just  how we can help you with the sort of strategy that was discussed in this post, contact us for more information.

12 Responses to “Modern Portfolio Theory turned upside down”

  1. Luke says:

    So much of this is misleading. First off, one should never use a time horizon of 10 years (and 10 cherry-picked years at that) as a representative sample size. If we take the 30 yr CAGR, the s&p had an approx. 7% annualized return (yes, that’s CAGR). When one considers that a simple .5% difference in return over that time period could impact oneself by about 25% in aggregate it is simple to see why one would prefer the market over a whole life policy, and I haven’t even brought up liquidity and taxes in this scenario.

    I’m also curious as to what the asset-mix of your representative “diversified portfolio” is in your example? I don’t know off hand but I’m guessing you chose the 70/30 mix, maybe 60/40? Would someone retiring in 2008 have had a 70/30 mix if working with an actual financial planner? No, they would have been far less exposed to the market.

    I get what you’re saying here and I’m not entirely opposed to insurance as potential alternatives but please be fair when trying to compare the two.

    • Brandon Roberts says:


      Define “in aggregate.”

      There’s nothing misleading here, and there are no rules that state one cannot nor should not evaluate S&P or diversified portfolio returns over a 10 year period. Would one not do so if he or she had a 10 year horizon?

      Furthermore, your juxtaposition of the last 30 years against the idea of assumable market returns too carelessly ignores the fundamentals of Steady State Economic Theory. While I completely agree that the growth of the U.S. stock market does not have a direct correlation with the U.S. economy (though strong evidence indicates a lagged covariance of the latter with the former) the entire U.S. economy is an underlying force. Unless you want to prescribe to rogue economic theory, you’re violating certain axiomatic notions by not appearing to accept the idea that the function of U.S. economic growth is a convergent sequence. This is the same economic theory that has predicted the slowing of U.S. economic growth and China’s catching up to U.S. economic achievement, which has been the correct trend.

      So, if we wish to make predictions about the next 30 years, we actually need to think in terms of the second order derivative of the stock market return where (I’m over simplifying) the function is return given time.

      Now the good news for those who hawk stocks is the expected return for stocks will still–most likely–outperform fixed assets (and that’s the category in which cash value life insurance sits). Meaning, this isn’t an us vs. them debate.

      Until you have to make one very important decision. If the expected value of returns across the board begins to decline due to the limit, it’s not likely to be a linear decline across asset classes. This means there’s another limit we must deal with concerning the risk premium. And you and I can preach till we’re blue in the face the worthiness of assuming the higher risk. The actual taking part in that is entirely dependent on perceived value.

      Understand that we believe stocks are all about the trade. And as someone who has made a fair amount of money trading stocks and their derivatives, I like that position. But, we can make out quite well, with less exposure to risk accepting a higher percentage of our portfolio sitting in fixed assets, and we can yield quite well when we use things like blended/over-funded whole life insurance and/or guideline premium funded universal life insurance.

  2. Luke says:

    Who is saying that this “diversified portfolio” is purely U.S. stocks, regardless of whether one accepts steady state or not?

    • Brandon Roberts says:

      I didn’t. But if you want to lean on the S&P 500’s 30 year historical yield as evidence that yield projections are off, you have to be very willing to evaluate how well U.S. stocks will perform in the future.

  3. Luke says:

    The 25% number is using, admittedly, simple math by comparing one portfolio earning an average of .5% greater than an alternative over 35 +/- years. If person A ended up with $1m at end of said time period, person B ended up with $750k at end of said time period. Before you castigate me over arithmetic/geometric returns and how we don’t know what the sequence of returns in this scenario would be, let me state that I understand that it won’t definitively reduce aggregate returns be 25% but rather it will be reduce aggregate returns by approx. 25%.

    As far as a 10 yr time horizon….unless one is planning for a single, 75 yr old retiree then no, one should not use a 10 yr time horizon (or possibly savings for a single, specific purchase). Keep in mind that if one reaches retirement at age 65 and is married, he/she has a 50/50 likelihood of reaching age 92. Oh, and this is using data of those who have already passed, which means that our estimates are likely on the modest side.

    I would also like to point out that the 7% CAGR number I noted was inflation adjusted at that.

    • Brandon Roberts says:

      You missed the point about thinking in terms of the second order derivative, and that’s precisely what looking at the last 10 years vs. the last 30+ accomplishes.

      Now, ideally if we believe that the underlying data in its aggregate represents random chance (i.e. in any given year there’s an equal probability of any return observed from systematic random sampling) than we wouldn’t simply take any time period, but rather take the parameters of the data characteristics within a time period and use those characteristics to generate random data from which we’d build a model in monte carlo to make inferences about the anticipated probability of returns; I’ve done such a thing in the past.

      You’re suggestion that the last 10 years or any 10 years is subject to selection bias (or “cherry picking” as you so pejoratively put it) can just as easily be thrown right back at you for your decision to select results from the last 30 years. But the beauty in our approach is that it at least acknowledges strong economic theory that would suggest the next 30 years might not be as great as the last (i.e. the rate of change moving forward is likely to be less than the rate of change from the past).

      And since we’re talking about the enriched or impoverished life of a person, is it not then prudent to assume low? You can’t take data from results that worked out the way you intended and suggest they support your theory unless you can point to a theory that supports why it worked. Getting it “correct” can at times simply be a function of two non related items matching up by chance. There are no economic theories that support the notion that the U.S. stock market will yield as it did or better than it did over the last 30 years.

      What you have to understand is how perfectly (though crudely) the 10 years evaluation takes an important economic theory into account.

      As far as the inflation adjusted CAGR goes, again we move the start and end dates to get wildly different results. You only get once chance to get this correct, so it’s best to challenge the idea and evaluate the minimum and maximum results, and plan expectations accordingly.

      Further, when it comes to the quoting the results of the S&P 500 or any index one must remember that an index can, over time, do something an investor cannot. Indexes drop under performing stocks in favor of others. This adjustment to who is in and who is out makes for an awfully tough time matching exactly what the index does over time. This is evident from the lagging results we see from some index funds. In other words, the market doesn’t much care how it performs over the long run. The individual investor very much does care.

      • Luke says:

        And you missed the point that a properly constructed portfolio would not be comprised of solely domestic equities…and again, one must accept your theories. Back to stable state, it’s been called many times in the past but of course we seem to forget those times. It’s analogous to every generation believing they are living in the end times….anyway, your second derivative analysis merely tells us what has happened and only tells us what CAN happen when using one’s interpretation. That is of course assuming we trust the inputs of the analyst. Needless to say, I personally don’t put much stock (see what I did there, lol) into it.

        If you don’t like my choice of 30 time horizon then by all means pick any other 30 yr time horizon, your choice (all that being said, I would even suggest that a 30 yr horizon is still too short for the majority of investors, but alas, that doesn’t really matter. The returns appear to deviate back to the mean once one achieves a large enough sample size, to my knowledge around 40 +/- yrs.). However, it doesn’t matter which period you choose, it will beat your whole life policy returns rather vigorously. So yes, I agree, you only get one chance, one should make the most of it and not leave returns on the table.

        And yes, one can do what an index does. Regardless, even if one cannot perfectly replicate index returns one can very closely match them. The Spartan 500 Index has lagged the S&P 500 index by an annualized .19%, attributed to mgmt fees, although not ideal it is hardly crippling.

        • Brandon Roberts says:

          But you see, you keep moving the goal post. First you contend (incorrectly) that I’m forgetting that a “properly constructed portfolio” is not constructed solely of domestic equities, and yet you continue to talk about nothing but equities. Then you name drop an index fund.

          As far as “its beating [my] whole life policy,” I’d ask you to go back to my original reply to you where I very clearly stated “this isn’t an us vs. them debate.” I believe I even pointed out that we should expect equities to yield higher. Now again, I’m not forgetting the whole properly diversified portfolio business as your comment was made against the backdrop of what time period one analyzes for S&P 500 returns.

          Steady state (not stable state) economics as far as I’m discussing it, has nothing to do with the allusion to end times. It merely notes that as economies mature they grow less rapidly. And from there we can theorize that a slower growing economy will decelerate the expansion of the stock market.

          You don’t place much stock in my model because you’ve shown no identifiable ability to understand it. I’ve even made some intentionally awkward comments that could be interpreted in a manner that you should have been more than capable of calling out, but instead you bring only cursory arguments to the table that reiterate the same sales jargon we’ve seen plenty of times before.

          I’ll help you with your counter argument. You should have jumped on the notion that the introduction of an exogenous variable such as new technology could accelerate economic growth. This is exactly what we saw in the 1980’s through the 1990’s with the introduction of personal computers. Now, there is a problem with what exactly that new technology would be, but one could hold out hope that it will eventually arrive.

          So again, it’s not about the fact that people own stocks, it’s how much they are told to own that can be a problem. And since you’re so concerned about the proper portfolio, why don’t you give us details?

          • Luke says:

            Am I moving the goal post or am I correctly reinforcing my original assertion that this write-up is highly misleading? Skirting the point that a larger sample size is more adequate than a smaller sample size doesn’t make you right, it just means you’re skirting.

            As to whether this is an “us vs. them” debate. After reading several posts on this site it is clear that you and your colleague are implying that it IS an “us vs. them” debate. So I’m welcoming it. By the way, there is only one of us who is “hawking” (as you so pejoratively put it) something and it isn’t me. One of us is entitled to a handsome commission by selling a product and the other is not in any sort of sales capacity.

            Stable State/Steady State, freudian slip (are you going to admonish me on my grammar skills because I didn’t capitalize Freudian there?). As to whether I understand any theory or all theories stated here, would you really like me to resort to the sort of ego-stroking vernacular usage that is ever so prevalent on this blog? Shall we discuss Pareto Optimality and it being a natural precursor to MPT (ya know, we could discuss this so that our readers can think to themselves, “wow, this guy is using terms I’ve never heard, he must know what he is talking about) or perhaps the New Zealanders monetary experiment with the Taylor Rule?

            Did I not jump on the fact that steady state (oh, there I typed it correctly, I now know what I’m talking about) is an old theory that can never be disproven as time is infinite and those spouting the merits of it can merely say, ad infinitum, that we are likely entering it. End times believers, as I clearly stated, were ANALOGOUS to steady state believers….analogous….not equated.

            As to the index used, what else is there to say? I provided you with an all world option since you didn’t like the S&P 500 due our, largely uncorrelated, economy supposedly reaching peak levels.

          • Brandon Roberts says:

            I’ve already explained the theory behind the chosen date range and you’ve said nothing to refute the idea outside of incorrectly depicting the intent behind the economic theory at play, and completely ignoring the second aspect to it. If you don’t agree with the assumptions you simply don’t agree with the assumptions, but you can’t continuously declare it misleading because you don’t agree with it. Instead you’d simply propose a reason as to why it’s fundamentally wrong, something you’ve not yet done. Hint: it requires something a tad more mentally taxing than simply looking at a historical chart and saying “well that must be what’s going to happen, again.”

            Are you sure the sample size upon which you base your assertion that this blog only seeks to create an US vs. THEM debate is sufficiently large enough? You seem to have deep concerns about sampling bias. With respect to your impertinent suggestion about the way in which we monetize with our business model, I’ll take a moment to highlight the fact that we use this blog as a resource to educate just as much as we do to make sales. I’ve personally turned away a handful of people this year alone whom I thought were in situations that couldn’t be made better by what I do, and I’ve offered my thoughts and advice to dozens more for absolutely free.

            Further I’ll note that I allow you to come to my venue to carry on your incoherent blather for going on six comments now. I don’t delete or alter your comments in any form, and I’m willing to take time out of my day to address you, point by point I might add. I don’t know you, so I won’t make any asinine public statements about what I think of you personally or where your motivations are. I’ll respectfully ask you to extend the same courtesy to me.

            You don’t enter into steady state. It’s a theory about how economies behave as they age. Most specifically what happens as they mature. We don’t need to know what stage we are in, the notion that the growth function possesses diminishing output given time acts as our guiding principal.

            Now, because I care more about truth and having a thoughtful discussion, I’ll ask you again (that’s right, despite your combative and rude nature I’ll continue to have this discussion with you publicly for everyone to come and see): because this is a discussion about MPT, and you’ve now suggested numerous times that there’s some confusion on properly structured portfolios, please provide details on said portfolios (and I’ll even throw you a bone on this one, I’m intentionally asking this question incorrectly).

          • Luke says:

            There is no standardized set of rules for what a well constructed portfolio should be just as there isn’t a standardized set of rules for what a healthy diet should be. However, just because there aren’t set rules doesn’t mean we aren’t able to identify an unhealthy portfolio composition just as it isn’t hard to identify an unhealthy diet. I can only say that a generally accepted allocated portfolio would have international exposure.

            Now, you’ll have to excuse me as the second derivative of the noble hosts’ snarkiness is forcing me to decrease my exposure to this fine blog.

Leave a Reply


  • Archives

  • Categories