Efficient frontier?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
Dave C.
Posts: 301
Joined: Sun Jul 10, 2011 10:18 pm
Location: Mesa, AZ

Efficient frontier?

Post by Dave C. »

"Efficient frontier" is a term that is new to me and one I would like to get back some feedback from this form.

The term first popped up when I began to use the investment software program "personal capital". First of all, I'll say that my use of "personal capital" has been very satisfactory to me. I only use the free services, but it is well organized, and puts every aspect of my portfolios in clear view. It breaks down my overall asset allocation, tabulates my net worth, tracks my spending and even makes some limited recommendations to help me keep my asset allocation aligned with my overall investing philosophy.
Here is where the "efficient frontier" comes into play. The service tells me that my asset allocation, which is roughly 50% equity and 50% fixed income, could be a bit more conservative. Their computerized analysis places us about 10% to "heavy" in equities, and about that much to light in fixed income vehicles.
In order to provide some context, I am 64, married, retired for three years with the portfolio of about $1.6 million.
90% of our holdings are tax protected, we are living comfortably off the taxable money and my pension. We'll both start SS at 66 to the tune of about $3400/month combined. $5,000 a month living expenses suits us very well.
I think that "efficient frontier" seems to be a modern-day investment term, right?
Does the calculation of the term make sense to some Bogleheads? Does anyone use it?
"Personal capital" seems to be telling me that if I have won the investment war, why not reduce risk?
They are making no specific recommendation for buys.... So do you guys think that the "efficient frontier" is a viable tool for measuring risk versus gain?
Thanks.
Easy does it/Live & Let Live/One day at a time. Thanks Bill.
JoinToday
Posts: 1110
Joined: Sat Mar 10, 2007 8:59 pm

Re: Efficient frontier?

Post by JoinToday »

Dave C. wrote:.....
Here is where the "efficient frontier" comes into play. The service tells me that my asset allocation, which is roughly 50% equity and 50% fixed income, could be a bit more conservative. Their computerized analysis places us about 10% to "heavy" in equities, and about that much to light in fixed income vehicles.
In order to provide some context, I am 64, married,....

Does the calculation of the term make sense to some Bogleheads? Does anyone use it?
"Personal capital" seems to be telling me that if I have won the investment war, why not reduce risk?
They are making no specific recommendation for buys.... So do you guys think that the "efficient frontier" is a viable tool for measuring risk versus gain?
Thanks.
First off, when I first learned about the efficient frontier, I thought it was fascinating. Optimizing your asset allocation to improve returns and reduce risk. Or improve returns at a given risk. The problem is that no one knows the correlation between different asset classes in the future, so it is of limited or no benefit.

Secondly, read books by WIlliam Bernstein. I think he writes about the efficient frontier. The Four Pillars of Investing. Larry Swedroe has the philosophy of if you have already won the game, why continue to play. Read his books also.

Third: I bet the program you mentioned can give your the "correct asset allocation" to about 15 digit of resolution, maybe more ..... with error bars (uncertainty) of +/- 30%. I wouldn't pay any attention to their recommendation. Zero.

From where you are, 40% to 60% equity is good. Even 30% equity should be OK. Nobody knows the future. Just pick a reasonable AA, and stay the course, and don't look back. I am in my late 50's, retired. 60% equity, 40% bonds. There was some nobel prize winner (I can't remember who) in finance who has a 50%/50% asset allocation. (Eugene Fama? French?). You are good where you are.
I wish I had learned about index funds 25 years ago
Topic Author
Dave C.
Posts: 301
Joined: Sun Jul 10, 2011 10:18 pm
Location: Mesa, AZ

Re: Efficient frontier?

Post by Dave C. »

This is the O.P. and I thank you very much for your quick reply. Each point you made makes all the sense in the world.
Easy does it/Live & Let Live/One day at a time. Thanks Bill.
User avatar
Epsilon Delta
Posts: 8090
Joined: Thu Apr 28, 2011 7:00 pm

Re: Efficient frontier?

Post by Epsilon Delta »

Dave C. wrote:"Efficient frontier" is a term that is new to me and one I would like to get back some feedback from this form.

...

I think that "efficient frontier" seems to be a modern-day investment term, right?
"Efficient frontier" is a term from multi-variable optimization, a branch of mathematics. Markovitz's Modern Portfolio Theory is an example of multi-variable optimization, so it uses that math, but there are many other fields that use the same math. "Supply chain optimization" is but one example.

Loosely, the efficient frontier is the set of all possibilities that are not obviously silly. For example if we assume everybody likes all kinds of fruit and give people a choice of:
  • 1 apple and 2 oranges
  • 1 apple and 1 orange
  • 2 apples and 1 orange
We will find that people who like oranges more than apples make the first choice while people who like apples more than oranges make the third choice while nobody chooses the second option, that is the efficient frontier is the first and third option, while "1 apple and 1 orange" is not on the efficient frontier.

Dave C. wrote: Here is where the "efficient frontier" comes into play. The service tells me that my asset allocation, which is roughly 50% equity and 50% fixed income, could be a bit more conservative. Their computerized analysis places us about 10% to "heavy" in equities, and about that much to light in fixed income vehicles.
The efficient frontier does not come into play here. If you're just looking at equities and fixed income it is very hard to come up with a realistic case where 40:60 is on the efficient frontier while 50:50 is not. Because both portfolios are on the efficient frontier you cannot prefer 40:60 because it is on the efficient frontier. You should prefer 40:60 if it suits your personal risk tolerance. Other people with a different risk tolerance will choose the 50:50 portfolio. Neither choice is inefficient, or obviously silly.

If the software is using "efficient frontier" to say you should prefer one of these allocations over the other it is bamboozling with buzzwords and should be regarded with skepticism.
User avatar
nisiprius
Advisory Board
Posts: 52212
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Efficient frontier?

Post by nisiprius »

The "efficient frontier" is language coined by Harry Markowitz as part of what's now called "modern portfolio theory." As typically presented nowadays, it incorporates ideas from William F. Sharpe and the "CAPM" (Capital Assets Pricing Model).

As a way of thinking about investment portfolios, it is important. As something to be used in a practical way in real life, it is... mostly hogwash, and is used by advisors as a "secret miracle ingredient" without much explanation. A good example of misuse would be a statement like this (possibly a straw man, this isn't an actual quotation from anybody). "If you want less risk in your portfolio, you could do that by adding bonds, but that would cut your return. A better way, which we know how to do, is to reduce risk by applying modern portfolio theory, and selecting assets with low correlation."

Read this carefully. This a simple explanation of the question that MPT answers:

BEGIN:

Let's suppose you have a set of assets... perhaps "U.S. stocks," "international stocks," and "bonds." Let's suppose you have records of the return of each of those asset classes, annually from 1926 to the present. From these records, you can calculate: the average annual return of each asset; the standard deviation (variation) of each asset; and the cross-correlations between the assets. These are the data that will go into the MPT calculation. To anticipate... they are the "garbage in."

Now, let's suppose you have a wheel of fortune labelled 1926 through 2015, and you have an absolute guarantee that 2016 will be exactly the same as one of the years 1926 through 2015, but you don't know which. You will spin the the wheel, and if it comes up "1958" then U.S. stocks, international stocks, and bonds will each have the same return they had in 1958.

Imagine doing this repeatedly on some specific portfolio--say, 40% U.S. stocks, 10%, international, and 50% bonds--and doing statistics on the results. Over a long period of time the returns from the portfolio will have an average return and an average standard deviation. Using CAPM you calculate the "Sharpe ratio," which is a measure of risk-adjusted return.

MPT tells you which portfolio will have the best risk-adjusted return... if you are investing in spins of this wheel of fortune. That is to say, if you are given the opportunity to bet repeatedly on spins of this wheel of fortune, MPT tells you the optimum portfolio on which to bet.

END.

The big problem--quite apart from philosophical issues about what kind of "risk" we care about, etc.--is that in reality, the returns from future years don't even come close to matching the model of "exact replays of randomly chosen years of past history."

To put it another way, all of the numbers we need to have to crank into the MPT model are just not known accurately. For example, we don't know the future correlations of any pair of assets--they are even less stable than the future return of any asset. MPT is garbage in, garbage out. And the numbers available to us to feed MPT are garbage.

In real life, just about everyone agrees that MPT, if used in a straightforward way, yields results that are "nonsense." Typically, the optimum portfolio changes wildly from year to year depending on the range of years fed into it, and amounts to not much more than a recommendation to bit really big on whatever asset class did the best.

Organizations that claim to be following MPT will, on closer inspection, say that they are following the "Black-Litterman" method, which includes predictions, estimates, or guesses about how asset classes will perform going forward. So, it's "garbage in" in a different kind of way. It amounts to saying if we knew some statistical characteristics about asset behavior going forward, then this would be the optimal portfolio. But, of course, we don't know.

Let me give you a concrete example. REITs are often touted for use in MPT-informed portfolios because they "have" similar return and standard deviation to stocks but "have" negative correlation with stocks. For example, during the three-year period 2000, 2001, and 2002, when the stock market had a miserable decline, REITS had a very solid gain, and a portfolio with both REITS and stocks would have enjoyed a much smoother ride. Unfortunately... despite that "low correlation" (or perhaps because of it, because low correlation doesn't mean negative correlation), REITS crashed worse than stocks in general during 2008-2009 and increased risk rather than moderating it. In short, it is basically a lie to say that REITS "have" low correlation with stocks. As with everything else, all we can say is that in the past they "have had" low correlations.
Last edited by nisiprius on Thu Apr 07, 2016 8:42 am, edited 2 times in total.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: Efficient frontier?

Post by dbr »

For another practical example, sign up for Financial Engines and follow its directions for modifying your portfolio every few months.
User avatar
quantAndHold
Posts: 10141
Joined: Thu Sep 17, 2015 10:39 pm
Location: West Coast

Re: Efficient frontier?

Post by quantAndHold »

The Personal Capital app is really cool, how it can slice and dice, and give you practically every stat you ever wanted about your portfolio. But the automated suggestions...lets just say...it's always complaining about something. I think it's a plot to get you to use their service. If I could turn off the suggestions, I would.
notsobright
Posts: 77
Joined: Thu Jun 28, 2012 9:07 pm

Re: Efficient frontier?

Post by notsobright »

I think it helps to see a graph.

http://www.spreadsheetml.com/finance/im ... ontier.png

Efficient frontier usually compares 2 investments going from 0% to 100%.

It plots average return vs volatility at all the various asset allocation mixes 0/100 , 10/90, 50/50...90/10... 100/0

Then theoretically, you can simply pick the hump out of the curve to get the "Most bang for your buck"
michaelsieg
Posts: 625
Joined: Mon Jan 07, 2013 10:02 am

Re: Efficient frontier?

Post by michaelsieg »

I agree that the efficient frontier is not that important for the OP.

The key question for him is to have a careful look at his risk tolerance. I think it really matters to look at the maximal drawdown of a portfolio if equity markets have a bear market. There was a good discussion on this topic in February last year:

Best AA if you've won the game viewtopic.php?t=165796

In this thread, there is a good link to an article by Rick Ferry about the AA for retirees, looking at outcomes during the last financial crisis
http://www.rickferri.com/blog/investmen ... -retirees/

Also, somewhere in the wiki (can't find it now) there is a simple table that looks at maximal drawdown of a portfolio in correlation with the equity/bond mix.
User avatar
Rx 4 investing
Posts: 735
Joined: Sat Apr 25, 2015 11:03 am

Re: Efficient frontier?

Post by Rx 4 investing »

DaveC wrote: “…The service tells me that my asset allocation, which is roughly 50% equity and 50% fixed income, could be a bit more conservative. Their computerized analysis places us about 10% to "heavy" in equities, and about that much to light in fixed income vehicles.”
Fellow Boglehead Rick Ferri has proposed a 30% stocks / 70% bond portfolio is the right place on the “efficient frontier” for pre-retirees and retirees. Ferri wrote this in his Forbes article:
“I propose the center of gravity for those who have accumulated enough for retirement to be 30% stocks and 70% bonds. This is a conservative mix that has enough equity to growth with inflation and enough fixed income to keep portfolio volatility at bay. Historically, a 30/70 allocation has earned the highest Sharpe ratio. This is the point on the efficient frontier that has earned the best risk-adjusted return”
http://www.forbes.com/sites/rickferri/2 ... fd34e7ac24


Here’s a brief passage from Mr. Bogle’s book The Little Book on Common Sense Investing, Chapter 18, “What Should I Do Now?”. I underlined the key part of the passage from Mr. Bogle’s writing that will probably apply to most pre-retirees and retirees.

“…Most of us will want more stocks when we’re young, have relatively small assets at stake, many years to recoup losses, and do not depend on investment income. When we’re older, we’re likely to prefer more bonds. If we’ve planned intelligently and invested wisely, our asset accumulations have grown to substantial size; we have far less time on our side; and when we have retired we will rely on our portfolios to produce a steady and continuing stream of income…”

The theory behind the complex investing concepts has been covered by previous posters. And as correctly pointed out, there is uncertainty in all of it. It turns out the portfolio with the least amount of uncertainty is a portfolio with a “duration” that matches the investment horizon.

The Price / Dividend ratio offers a pretty decent approximation of the modified duration of the S & P 500 index.

--Current S & P500 price here: $2,064 http://www.multpl.com/s-p-500-historical-prices

--Current dividend here: $43.50 http://www.multpl.com/s-p-500-dividend/

--The modified duration of the S & P 500: 47 years (Stocks = S)

-The average duration of Vanguard’s Total Bond Market ETF (BND): 5.6 years (Bonds = B)

Find that here: https://personal.vanguard.com/us/funds/ ... =INT#tab=2

--The duration of cash: zero (0) years (Cash = C)

Here’s the current approximate duration of typical 50%-S /40%-B /10%-C portfolio:

(0.5 x 47) + (0.4 x 5.6) + (0.1 x 0) = 23.5 + 2.24 + 0 = 25.7 years.

Questions:
1. Are you looking for some degree of predictability for money that needs to be spent in the future?
2. Does most of the money committed to your 50/50 portfolio have a time horizon of 26 years?


Investment expert William Bernstein wrote this cautionary note a few years back about market “duration” and retirees…

“… equanimity to market declines depends on time horizon. If you’re retired and living off savings, you will neither have enough time to get over the duration hump nor be able to make the contributions to shorten it.”

As mentioned, if a buy-and-hold investor with no particular view about market conditions or future returns wishes to have a fairly predictable amount of wealth at some future date, that investor should hold a portfolio with a duration that is roughly equal to the investment horizon. As you can see in the current 50/50 example calculation, the number carrying the most weight is the stock allocation. Thus, a risk-averse investor should conclude that any money needed in the next 5-10 years would not have a large allocation to US stocks.

You can do a quick exercise to determine your comfort level with your current 50/50 allocation using a simple mental accounting approach like “bucketing. ”

--Start by labeling your various buckets of money--- with a time horizon--- of when that money needs to be available for spending.

--You can then use the duration numbers of stocks, bonds, and cash to design portfolio allocations that match duration to the time horizon of each of your buckets.

--Should you find this approach more personalized to your situation than the "canned" allocation advice out there, the calculations can then be done annually to ensure your various portfolio’s duration and time horizons stay aligned.

Hope this provides some additional perspective. Good luck and continued best wishes for a comfortable retirement. :beer
Last edited by Rx 4 investing on Fri Apr 08, 2016 7:31 pm, edited 1 time in total.
“Everyone is a disciplined, long-term investor until the market goes down.” – Steve Forbes
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Efficient frontier?

Post by patrick013 »

Whether this results in a 30/70 portfolio I don't know, but it's something
that can be done with a Buffett portfolio. Main question is...how long does
the recession last ?

Have enough bonds in an Interm Trsy Fund to last thru the longest recession.
Spend those flight-to-quality cap gains during a stock decline or major lengthy
recession. Spend the whole Trsy fund if necessary waiting for the stock fund
to recover.

Otherwise, keep your S&P 500 investment at whatever ratio occurs then and spend
that down thru cap gains in non-recession times where some price appreciation and
total returns are experienced and recovery from a possible recession has occurred
and new gains are there to withdraw and spend.

Thoughts on this strategy using a bond reserve to weather a full recession ?
age in bonds, buy-and-hold, 10 year business cycle
TomCat96
Posts: 1040
Joined: Sun Oct 18, 2015 12:18 pm

Re: Efficient frontier?

Post by TomCat96 »

Dave C. wrote:"Efficient frontier" is a term that is new to me and one I would like to get back some feedback from this form.

The term first popped up when I began to use the investment software program "personal capital". First of all, I'll say that my use of "personal capital" has been very satisfactory to me. I only use the free services, but it is well organized, and puts every aspect of my portfolios in clear view. It breaks down my overall asset allocation, tabulates my net worth, tracks my spending and even makes some limited recommendations to help me keep my asset allocation aligned with my overall investing philosophy.
Here is where the "efficient frontier" comes into play. The service tells me that my asset allocation, which is roughly 50% equity and 50% fixed income, could be a bit more conservative. Their computerized analysis places us about 10% to "heavy" in equities, and about that much to light in fixed income vehicles.
In order to provide some context, I am 64, married, retired for three years with the portfolio of about $1.6 million.
90% of our holdings are tax protected, we are living comfortably off the taxable money and my pension. We'll both start SS at 66 to the tune of about $3400/month combined. $5,000 a month living expenses suits us very well.
I think that "efficient frontier" seems to be a modern-day investment term, right?
Does the calculation of the term make sense to some Bogleheads? Does anyone use it?
"Personal capital" seems to be telling me that if I have won the investment war, why not reduce risk?
They are making no specific recommendation for buys.... So do you guys think that the "efficient frontier" is a viable tool for measuring risk versus gain?
Thanks.

I don't believe this is correct. I could have an investment allocation that is 100% Treasuries and still be on the efficient frontier.
Likewise, I could be 100% stocks and still be on the efficient frontier.

The efficient frontier is a graph of all possible investments which return the highest return for the least amount of risk.
For example:

Your bank sells bonds at 2% interest.
Treasuries are paying 2% interest.

The treasuries are on the efficient frontier, but the bank bonds are not.
The idea here is that the bank is riskier than the US government. The bank could go down, could have a run, etc. The US government on the other hand is more credit worthy. If the bank just so happens to go down, the US government will bail them out anyway. (Via conservatorship/FDIC ins.)

So if you were going to invest in an instrument paying 2%, you want to pick the one with the least amount of risk for that 2%.

That's the idea in a nutshell.


As a day to day practical tool, it's useful because it makes you ask the question:
Am I assuming a proper amount of risk for my expected returns?

If a well diversified portfolio of stocks (ex. VFINX) is paying 10% per year, and my uncle chuck, the alcoholic, wants to borrow money from me paying 10% a year, then the conceptual tool of the efficient frontier allows me to recognize that uncle chuck offers an inferior investment.
There is too much risk for the same amount of return.


So generally speaking we know of two points on the efficient frontier.
1. US treasuries, considered risk free.
2. The market. VFINX or vanguard total market.

But, what we cannot be sure of what is the proper ratio of allocation that is also efficient. Whether 30/70, 50/50, 60/40 are efficient is anyone's guess. In fact, they may well all be efficient.

For example, suppose I am looking for a 9% rate of return. How can I achieve that with the least possible risk.
Well, the market is the point for 10%. So the optimal way to achieve that is probably some stocks mixed with some lower return, but less risky asset.
That less risky asset could be bonds, it could be treasuries, it could any number of the plethora of possible investments.

It could be 99% stocks and 1% treasuries.
It could be 90% stocks 8% commercial paper, 2% treasuries.

Who knows?
But at that point, the conceptual tool of the efficient frontier ceases to be of practical use for investing.
Last edited by TomCat96 on Thu Apr 07, 2016 1:57 pm, edited 1 time in total.
User avatar
Rx 4 investing
Posts: 735
Joined: Sat Apr 25, 2015 11:03 am

Re: Efficient frontier?

Post by Rx 4 investing »

patrick013 wrote: "Thoughts on this strategy using a bond reserve to weather a full recession ?"
I think your treasury bond allocation approach has merit, but for higher stock allocations, 100% intermediate term treasuries are not likely to appreciate enough to mitigate the kind of drawdown like we saw in '08-'09.

If you're speaking about the Buffett heirs' 90% SPY / 10% bond portfolio, I would consider allocating a healthy portion of that treasury allocation ( as much as 25%-50% ) to longer- duration treasuries. Rationale...

--In the '08-09 meltdown, the S & P 500 fell about -38% , while intermediate term treasuries rose about 18-20%.

--Comparably, iShares 20+ Year Treasury Bond (TLT) rose +31%.

That type of "flight to quality" appreciation from longer-duration treasuries can provide a very nice reservoir to draw upon for re-balancing during sharp stock market downturns.
“Everyone is a disciplined, long-term investor until the market goes down.” – Steve Forbes
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Efficient frontier?

Post by patrick013 »

Rx 4 investing wrote:
patrick013 wrote: "Thoughts on this strategy using a bond reserve to weather a full recession ?"
I think your treasury bond allocation approach has merit, but for higher stock allocations, 100% intermediate term treasuries are not likely to appreciate enough to mitigate the kind of drawdown like we saw in '08-'09.

That type of "flight to quality" appreciation from longer-duration treasuries can provide a very nice reservoir to draw upon for re-balancing during sharp stock market downturns.
The below chart would show that in 2008 TRSY withdrawals are desired.
Even in 2009 TRSY withdrawals are more desirable from 2008 gains still
available in TRSY fund(s). Having 3-5 years living expenses in TRSY bond
fund reserve would have worked out fine.

Image
age in bonds, buy-and-hold, 10 year business cycle
User avatar
nisiprius
Advisory Board
Posts: 52212
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Efficient frontier?

Post by nisiprius »

TomCat96 wrote:...I don't believe this is correct. I could have an investment allocation that is 100% Treasuries and still be on the efficient frontier.
Likewise, I could be 100% stocks and still be on the efficient frontier.

The efficient frontier is a graph of all possible investments which return the highest return for the least amount of risk...
Yes and no. Where CAPM and Sharpe come in is with the business of the tangent line. In effect you add a third asset to the mix, the "riskless asset." That's the dot on the vertical axis, representing 0% standard deviation, and hopefully some positive number (currently 0.25%) for the return from a riskless asset.

Straight lines from the riskless asset to any portfolio represent all of the results achievable by mixing in the riskless asset.

When you do this, 0% stocks and 100% stocks are no longer optimum, because, for any desired value of risk, the tangent line has higher return. That's what results in the specification of a single "optimum" value. In effect, it's the optimum portfolio for an investor who has the option of keeping some of their portfolio in cash.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
Taylor Larimore
Posts: 32842
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Re: Efficient frontier?

Post by Taylor Larimore »

Dave:

This is Investopedia's definition of "Efficient Frontier":
The efficient frontier is the set of optimal portfolios that offers the highest expected return for a defined level of risk or the lowest risk for a given level of expected return. Portfolios that lie below the efficient frontier are sub-optimal, because they do not provide enough return for the level of risk. Portfolios that cluster to the right of the efficient frontier are also sub-optimal, because they have a higher level of risk for the defined rate of return.
In my opinion, the "Efficient Frontier" has little or nothing to do with your own best stock/bond allocation which should be based on your goals, time-frame, risk-tolerance and personal financial situation.

Inasmuch as you "have won the investment war," consider doing what I do and sleep like a baby:

Invest your savings that you cannot afford to lose in bonds (consider Total Bond Market). Invest the rest in stocks (consider U.S. total market index fund and Total Market International (20%-30%). It's that easy --especially when "90% of our holdings are tax protected."

The Three-Fund Portfolio

Keep investing simple (and low cost).

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
Northern Flicker
Posts: 15363
Joined: Fri Apr 10, 2015 12:29 am

Re: Efficient frontier?

Post by Northern Flicker »

I don't believe this is correct. I could have an investment allocation that is 100% Treasuries and still be on the efficient frontier.
That's probably not true. Most efficient frontiers of US stocks and bonds are shaped like the right branch of a hyperbola, and as bond allocation increases, you reach a point where further increases in bonds increase risk while decreasing returns. There is a point of minimum risk beyond which further reductions in equities would be made holding cash if you want an efficient portfolio.

Depending on the mix of assets, computed efficient frontiers can be very susceptible to sample bias, and don't take into account risks that did not materialize during the sample period even though they were real risks that could have materialized.
Last edited by Northern Flicker on Fri Apr 08, 2016 11:55 am, edited 1 time in total.
User avatar
Tyler9000
Posts: 740
Joined: Fri Aug 21, 2015 11:57 am

Re: Efficient frontier?

Post by Tyler9000 »

My biggest problem with most efficient frontier calculations is that they too often focus on tweaking a mathematically perfect combination of a small handful of assets while ignoring the fact that there are so many other combinations of different assets that might be way better than anything on that curve. I've been playing with ways to look at this problem, although it's an ongoing project and I don't claim it's a be-all-end-all analysis or anything. Still, anyone interested in this topic might also find the results interesting.

Image

The equivalent to the efficient frontier in this chart is the top/right edge of the cloud. You can read this for more info. There's also some discussion in another thread here: viewtopic.php?f=10&t=171019#p2827370

FWIW, I'm personally skeptical of most "modern portfolio theory" or "efficient frontier" claims by many robo advisors. Too often it feels more like marketing than substance.
Last edited by Tyler9000 on Thu Apr 07, 2016 9:33 pm, edited 5 times in total.
User avatar
patrick013
Posts: 3301
Joined: Mon Jul 13, 2015 7:49 pm

Re: Efficient frontier?

Post by patrick013 »

Efficient Frontier is as basic and essential and conceptual
as many other Financial Theories. Amongst those include
portfolio diversification, bond maturity diversification,
stock portfolio risk premium, and inflation adjustments for
statistics like GDP. Some theories prove concepts, others
become quite pragmatic.

I always remember a proverb: when Theory becomes Practice then
Science becomes Art.

As long as std. deviation doesn't get too high an asset allocation
can be developed with the EF as a starting point or guide.
Most index funds have std. deviations that are considered reasonable
so the asset allocation is mostly to capture stock market gains
to bonds, hence, the age-in-bonds concept.
age in bonds, buy-and-hold, 10 year business cycle
User avatar
Tyler9000
Posts: 740
Joined: Fri Aug 21, 2015 11:57 am

Re: Efficient frontier?

Post by Tyler9000 »

nisiprius wrote: Organizations that claim to be following MPT will, on closer inspection, say that they are following the "Black-Litterman" method, which includes predictions, estimates, or guesses about how asset classes will perform going forward. So, it's "garbage in" in a different kind of way. It amounts to saying if we knew some statistical characteristics about asset behavior going forward, then this would be the optimal portfolio. But, of course, we don't know.
This point is especially important. I learned about Black-Litterman when researching robo advisor methodology -- it's basically PhD speak for "educated guess". Some companies like Wealthfront also introduce their own proprietary "risk tolerance metric", and the individual assets in the pool of options may be limited by business factors other than what would truly make for the best portfolio. While they use terms that make it sound very scientific, always remember that's part of the marketing strategy.
Topic Author
Dave C.
Posts: 301
Joined: Sun Jul 10, 2011 10:18 pm
Location: Mesa, AZ

Re: Efficient frontier?

Post by Dave C. »

This is the OP, and I offer my sincere thanks to everyone who has responded to my inquiry about the " efficient frontier".
I never expected such a large volume of responses, nor the information provided in those responses. I actually understood enough of your responses that it helped me to know what my concerns should be, and where they should not be. Thank you all again.
Easy does it/Live & Let Live/One day at a time. Thanks Bill.
fortyofforty
Posts: 2083
Joined: Wed Mar 31, 2010 12:33 pm

Re: Efficient frontier?

Post by fortyofforty »

It seems that nobody can see where the Efficient Frontier will be, only where it was. The trick is trying to guess where it is, and behaving accordingly. Or, just don't worry too much about it (my personal philosophy), and choose your asset allocation based on risk tolerance.
Benjamin Buffett
Posts: 129
Joined: Sun Nov 25, 2018 1:20 pm
Location: Indiana

Re: Efficient frontier?

Post by Benjamin Buffett »

nisiprius wrote: Thu Apr 07, 2016 8:37 am The "efficient frontier" is language coined by Harry Markowitz as part of what's now called "modern portfolio theory." As typically presented nowadays, it incorporates ideas from William F. Sharpe and the "CAPM" (Capital Assets Pricing Model).

As a way of thinking about investment portfolios, it is important. As something to be used in a practical way in real life, it is... mostly hogwash, and is used by advisors as a "secret miracle ingredient" without much explanation. A good example of misuse would be a statement like this (possibly a straw man, this isn't an actual quotation from anybody). "If you want less risk in your portfolio, you could do that by adding bonds, but that would cut your return. A better way, which we know how to do, is to reduce risk by applying modern portfolio theory, and selecting assets with low correlation."

Read this carefully. This a simple explanation of the question that MPT answers:

BEGIN:

Let's suppose you have a set of assets... perhaps "U.S. stocks," "international stocks," and "bonds." Let's suppose you have records of the return of each of those asset classes, annually from 1926 to the present. From these records, you can calculate: the average annual return of each asset; the standard deviation (variation) of each asset; and the cross-correlations between the assets. These are the data that will go into the MPT calculation. To anticipate... they are the "garbage in."

Now, let's suppose you have a wheel of fortune labelled 1926 through 2015, and you have an absolute guarantee that 2016 will be exactly the same as one of the years 1926 through 2015, but you don't know which. You will spin the the wheel, and if it comes up "1958" then U.S. stocks, international stocks, and bonds will each have the same return they had in 1958.

Imagine doing this repeatedly on some specific portfolio--say, 40% U.S. stocks, 10%, international, and 50% bonds--and doing statistics on the results. Over a long period of time the returns from the portfolio will have an average return and an average standard deviation. Using CAPM you calculate the "Sharpe ratio," which is a measure of risk-adjusted return.

MPT tells you which portfolio will have the best risk-adjusted return... if you are investing in spins of this wheel of fortune. That is to say, if you are given the opportunity to bet repeatedly on spins of this wheel of fortune, MPT tells you the optimum portfolio on which to bet.

END.

The big problem--quite apart from philosophical issues about what kind of "risk" we care about, etc.--is that in reality, the returns from future years don't even come close to matching the model of "exact replays of randomly chosen years of past history."

To put it another way, all of the numbers we need to have to crank into the MPT model are just not known accurately. For example, we don't know the future correlations of any pair of assets--they are even less stable than the future return of any asset. MPT is garbage in, garbage out. And the numbers available to us to feed MPT are garbage.

In real life, just about everyone agrees that MPT, if used in a straightforward way, yields results that are "nonsense." Typically, the optimum portfolio changes wildly from year to year depending on the range of years fed into it, and amounts to not much more than a recommendation to bit really big on whatever asset class did the best.

Organizations that claim to be following MPT will, on closer inspection, say that they are following the "Black-Litterman" method, which includes predictions, estimates, or guesses about how asset classes will perform going forward. So, it's "garbage in" in a different kind of way. It amounts to saying if we knew some statistical characteristics about asset behavior going forward, then this would be the optimal portfolio. But, of course, we don't know.

Let me give you a concrete example. REITs are often touted for use in MPT-informed portfolios because they "have" similar return and standard deviation to stocks but "have" negative correlation with stocks. For example, during the three-year period 2000, 2001, and 2002, when the stock market had a miserable decline, REITS had a very solid gain, and a portfolio with both REITS and stocks would have enjoyed a much smoother ride. Unfortunately... despite that "low correlation" (or perhaps because of it, because low correlation doesn't mean negative correlation), REITS crashed worse than stocks in general during 2008-2009 and increased risk rather than moderating it. In short, it is basically a lie to say that REITS "have" low correlation with stocks. As with everything else, all we can say is that in the past they "have had" low correlations.
This is one of the best descriptions of how practically useful, or useless, Modern Portfolio Theory is. In finance and and economics classes, and from financial advisors in my past, discuss the magic importance of MPT, and how it informs the most logical choice of assets based on the statistical data of correlations and risk. I always wondered why its answers varied from the data, I just always thought I was choosing the wrong data, or entering it wrongly. When I read into it it seemed like it was dependent on knowledge and data that would require predicting the future........

The idea of leveraging up your low risk assets seems a bit dubious to me, and at the time I remember thinking "how do I really know that it is low risk?". Long term historical charts did not reassure me enough to really go farther with MPT.
User avatar
vineviz
Posts: 14921
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Efficient frontier?

Post by vineviz »

Benjamin Buffett wrote: Mon Jun 24, 2019 4:56 pm This is one of the best descriptions of how practically useful, or useless, Modern Portfolio Theory is.
If it were completely, or even mostly, accurate I'd probably agree with you.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Post Reply