What do you expect to achieve with your Tilt?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
AlohaJoe
Posts: 2334
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: What do you expect to achieve with your Tilt?

Post by AlohaJoe » Sun Mar 12, 2017 11:34 pm

sean.mcgrath wrote:But my main point is there is no sane way for me to compare two portfolio approaches if I don't have any quantification of risk. Do you quantify it at all? If so, how do you quantify the relative length & breadth of drawdown risk between two investment approaches?


Sure, I try to quantify it. Or, maybe "quantify" isn't quite the right word because it implies boiling it down to single, simple numbers. But I tend to prefer comparisons of drawdowns & income equivalents. I don't tilt because I expect higher returns (though if that happens I certainly won't complain); I tilt because I expect it will have fewer deep drawdowns and less tail risk.

For drawdowns I look at things like

https://portfoliocharts.com/portfolio/drawdowns/

For incoming I look at things like

https://portfoliocharts.com/portfolio/withdrawal-rates/

and

https://portfoliocharts.com/portfolio/r ... -spending/

There are a lot of lessons to learn from historical data but you also have to be willing to make (hopefully educated) guesses when it deciding how much they will apply to the future.

Theoretical
Posts: 1063
Joined: Tue Aug 19, 2014 10:09 pm

Re: What do you expect to achieve with your Tilt?

Post by Theoretical » Mon Mar 13, 2017 12:43 am

My target max drawdown is 35% with a greater balance between bond and stock than afforded by most young equity allocations. Tilting allows me greater access (by taking on the risk) to increase my exposure to the unquestionable diversification of quality bonds without having to use external leverage or play with bond futures. I also believe it allows for different and sometimes offsetting sources of return.

My portfolio is a modified Robert T set at around 60/40 rather than his 75/25.

My goal is to meet or exceed 75% VT and 25 BND with left tail risks closer to a 1/3 drawdown than a 1/2 drawdown.

User avatar
meowcat
Posts: 363
Joined: Wed May 09, 2012 5:46 am

Re: What do you expect to achieve with your Tilt?

Post by meowcat » Mon Mar 13, 2017 6:43 am

I tilt heavily to Small and mid (not value). It's worked quite well for me in the past. My expectations are.. well, I've learned not to expect too much as high expectations often end in disappointment.
More people should learn to tell their dollars where to go instead of asking them where they went. | -Roger Babson

User avatar
Aptenodytes
Posts: 3741
Joined: Tue Feb 08, 2011 8:39 pm

Re: What do you expect to achieve with your Tilt?

Post by Aptenodytes » Mon Mar 13, 2017 8:28 am

We all inherit the same market history, so if expected return means historic returns it is better to look them up than to ask us.

If expected returns means what do I expect from future returns, then the level of precision goes way down. Nowhere near the tenth of a percentage point.

I don't estimate my expected return mathematically. I expect my tilt to provide returns about equal to the total market with less volatility (it is a Larry portfolio). That's enough precision for me. Somewhere along the line I'm sure I calculated an expected return but that was for portfolio construction purposes, and I'm no longer doing that.

betablocker
Posts: 251
Joined: Mon Jan 11, 2016 1:26 pm

Re: What do you expect to achieve with your Tilt?

Post by betablocker » Mon Mar 13, 2017 9:08 am

Dirghatamas wrote:
willthrill81 wrote:And you say that we are the ones making the bet? :oops:

Considering that since 2000 large caps have had a real annual return of 2.27% while small caps have returned 5.88%, the 21st century doesn't seem to favor that prediction so far.

When people start saying "this time, it's different," I've learned to back away slowly.


I don't think you are seeing my position at all. My position is always "we don't have a clue" and be defensive. I am NOT predicting large will outperform small. Look at your response. Now replace "small" and "large" with US vs. International. Every week, there are threads by posters looking back at the last 20 or 30 year data on US vs. International and asking "why on earth would anyone invest outside the US". Perhaps there is a "factor" called US which explains good returns :shock: Somehow, people find that argument not respectable and argue for global diversification saying past performance is not a good predictor and if the market consensus is such and such, what reason do we have to expect all that to be not factored in..

There isn't a conceptual difference between say Country bias or sector bias or so called factors: they are all bets away from the market consensus.


Except that country bias has been shown through actual returns to be a bad bet while certain factors have been shown to be good bets. You should get more return for more risk. That's just economics. If you don't agree then how do explain that unsecured debt costs more than debt backed by collateral? Please convince the bank that holds my small business loan of this fact if you can. When private equity buys small companies, they demand way more return than if they bought into the S&P. This is only logical. As you say there a many factors and most of them are data mining or mean reverting but there are a handful that have been shown to work across countries, across time, across asset classes, etc. They also have behavioral or risk explanations. Your argument that all deviations are the same isn't supported. You have to willfully ignore the evidence.

Theoretical
Posts: 1063
Joined: Tue Aug 19, 2014 10:09 pm

Re: What do you expect to achieve with your Tilt?

Post by Theoretical » Mon Mar 13, 2017 10:35 am

Dirghatamas, why don't you have any bonds? Your arguments re: anomalies and not knowing what will outperform may be accurate for your stocks, assuming that float-adjusted cap weighting is the appropriate world stock metric. T-bills outperformed the US stock market for an entire decade 2000-2010 and there was an entire century (1800s) where bonds outperformed stocks.

You're making a bet that only one class of assets will outperform and then criticizing people that choose to make the same judgement call but within an asset class. That's why you're getting pushback.

Dirghatamas
Posts: 506
Joined: Fri Jan 01, 2016 6:18 pm

Re: What do you expect to achieve with your Tilt?

Post by Dirghatamas » Mon Mar 13, 2017 4:09 pm

betablocker wrote:
Dirghatamas wrote:
There isn't a conceptual difference between say Country bias or sector bias or so called factors: they are all bets away from the market consensus.

Except that country bias has been shown through actual returns to be a bad bet while certain factors have been shown to be good bets.


This is not true. There is 116 years of past data by Dimson/Marsh etc. showing the stocks returns on every major country/exchange. Even leaving aside wars and revolutions, there are vast differences between returns in each country. For the countries which did best (Australia, Sweden, US, South Africa) a buy and hold portfolio concentrated (tilted) to their country would have resulted in ~2% per year better returns. If you want to ignore very old history, you will find the similar data for say last 20 years. It may be politically incorrect but completely rational to ask if "country" is a factor. Influential people like Bogle, Buffett etc. certainly believe so and have described many such: rule of law, dynamic culture, capitalism ingrained in the culture..many others. Obviously, no academic will write scholarly articles on a narrative explaining these differences, but for a rational investor, there is no obvious reason why a tilting portfolio could not be created using such data. Indeed, most on this forum do that precisely by tilting to/away from their home country. It is just not one of the so called factors.

betablocker wrote: You should get more return for more risk. That's just economics. If you don't agree then how do explain that unsecured debt costs more than debt backed by collateral? Please convince the bank that holds my small business loan of this fact if you can. When private equity buys small companies, they demand way more return than if they bought into the S&P. This is only logical.


This is fair: risk premium for riskier assets. Again, there are many such ways to diversify riskier assets. Take the so called "emerging markets" which used to be called developing countries a long time back. Since roughly 1970, a portfolio heavily tilted to these emerging markets has indeed outperformed say the world market (but actual index funds for this class are much more recent). This fits the "risk" narrative of investors asking for a higher premium..but the reward stays while risk gets diversified away across so many countries. Indeed many people do tilt to emerging markets. But there are no factors labeled as such called "country risk". Why not? The total assets in Emerging markets are much larger than say small cap value so the chance of it being arbitraged away quickly are smaller (much bigger asset class). This again says there could be a factor called country.

Obviously, in all my posts on this thread, I am being somewhat sarcastic. I DON'T do any of these tilts. I am simply trying to point out to the tilters that there are dozens of possible tilts which could all be back tested to be stable and have excellent narratives on why they exist (but fail in future).

Dirghatamas
Posts: 506
Joined: Fri Jan 01, 2016 6:18 pm

Re: What do you expect to achieve with your Tilt?

Post by Dirghatamas » Mon Mar 13, 2017 4:24 pm

Theoretical wrote:Dirghatamas, why don't you have any bonds? Your arguments re: anomalies and not knowing what will outperform may be accurate for your stocks, assuming that float-adjusted cap weighting is the appropriate world stock metric. T-bills outperformed the US stock market for an entire decade 2000-2010 and there was an entire century (1800s) where bonds outperformed stocks.

You're making a bet that only one class of assets will outperform and then criticizing people that choose to make the same judgement call but within an asset class. That's why you're getting pushback.


This is completely fair. I fully accept that we are making bets. As you point out, my bet is that long term (not 10 years) stocks overall will outperform bonds. They may not. That's the bet I have always made.

What I am criticizing (if you read my OP in this thread) is the level of accuracy needed in trying to answer the question: how much do you expect to achieve with your tilt? Is 0.5% worth it? The reason I responded in the first place was to point out that we don't even know significant digits of say 20 year future returns, let alone decimal digits.

Investing is a highly imprecise activity. Creating very accurate and mathematical models may give an illusion of rigor, but the actual outcome may not even rhyme with your narrative, let alone look close.

In terms of what I do about forecasting, I don't. I accept that I will simply get the returns the stock market will create. I try to save a lot and model a SWR at far less than 1%. That is my margin of safety: invest in the historically best performing assets (stocks) but prepare for returns even worse than worst case of nominal bonds. I don't try to beat the market by tilting, because long term, beating the market is a very difficult thing to do.

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Tue Mar 14, 2017 1:31 pm

AlohaJoe wrote:Sure, I try to quantify it. Or, maybe "quantify" isn't quite the right word because it implies boiling it down to single, simple numbers.


Thanks, Joe! Funny enough I was just on that site before I posted, took a brief look at the drawdown, and thought "that's a dumb graph, I want the Std Dev." You are right: assuming that the site does the calculations properly, that is the information I actually want; it looks like it even lets me model buying on margin. I need to figure out how to get the sense of frequency out of the graph, but then I have pretty much a complete picture.

On a side note, "quantify" is exactly the right word for what you do. It will be a sad day when "quantify" means you're only allowed one number, and you'd better make it simple. :happy


dbr wrote:The OP suggested at least the expected mean and expected SD, the usefulness of SD being a point for discussion. The point is that all results are only one instance from a possible distribution and we plan portfolios based on the expected distribution. The whole point of the question, it seems to me, is to specify in numbers what the rationale is for any person's adoption of a tilted portfolio and what that portfolio actually is.


Theoretical wrote:My target max drawdown is 35%.

My portfolio is a modified Robert T set at around 60/40 rather than his 75/25.

My goal is to meet or exceed 75% VT and 25 BND with left tail risks closer to a 1/3 drawdown than a 1/2 drawdown.


By the dbr-improved question, Theoretical is now the first (and so far only) person to actually reply to my original post -- thanks for that! If you're ever in the Netherlands you get a free beer or something.


I will admit to being surprised at how few people try to understand in some sort of quantitative way what the "expect" to happen when they deviate from total market. I would recommend it: not to get one point behind the decimal, but to understand in a practical way what you are actually doing.

P.s. Dir, I'd hoped to avoid a discussion on whether Tilting actually works, as I was interested in learning about how others evaluate their risk / return trade-offs, but one point on the Tilters' theory: what they are saying is that these asset classes have an additional risk (and return) that is not correlated to the rest of the market. It doesn't make the asset classes special because they have a higher return, but because the return is an uncorrelated return. If that is true, then holding that asset class follows exactly the same logic as holding the entire market: that the market only rewards risks that cannot be diversified away.

Thanks all for the very interesting comments and discussion.

Cheers,
Sean
Last edited by sean.mcgrath on Tue Mar 14, 2017 3:29 pm, edited 1 time in total.

itstoomuch
Posts: 4454
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: What do you expect to achieve with your Tilt?

Post by itstoomuch » Tue Mar 14, 2017 1:58 pm

Deleted :sharebeer
Last edited by itstoomuch on Tue Mar 14, 2017 2:56 pm, edited 1 time in total.
Rev90517; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax 25%. Early SS. FundRatio (FR) >1.1 67/70yo

Dirghatamas
Posts: 506
Joined: Fri Jan 01, 2016 6:18 pm

Re: What do you expect to achieve with your Tilt?

Post by Dirghatamas » Tue Mar 14, 2017 2:22 pm

sean.mcgrath wrote:P.s. Dir, I'd hoped to avoid a discussion on whether Tilting actually works, as I was interested in learning about how others evaluate their risk / return trade-offs, but one point on the Tilters' theory: what they are saying is that these asset classes have an additional risk (and return) that is not correlated to the rest of the market. It doesn't make the asset classes special because they have a higher return, but because the return is an uncorrelated return. If that is true, then holding that asset class follows exactly the same logic as holding the entire market: that the market only rewards risks that cannot be diversified away.
Cheers,
Sean


Sean

First, I apologize for derailing your thread with posts arguing against tilting. I thought more, and these type of discussions rarely change people's views, so it is pointless to argue investing philosophy.

I have a more serious/ operational question in terms of learning from you guys. How do you guys actually implement this in terms of accumulation or withdrawals? If there are lets say 2 assets in one's portfolio and they zig and zag and are uncorrelated, then on paper this doesn't help you, unless you re balance (move money between them) in accumulation or selectively take money out of them in withdrawal. However, the compelling data for re balancing between at least stocks and intermediate/short/total bonds is that it doesn't lead to better end value. There is some data on re balancing bonus between long bonds and stocks but long bonds are frowned on this forum (correctly so due to very high risk to interest rate).

So, if you combine say reasonable duration bonds and tilted stocks (say small cap value or whatever), aren't you simply relying on the assertion that the tilt will lead to higher returns, because small cap value will outperform? If not, are you guys re balancing between different stock classes?

This time I am not trying to argue, simply trying to understand how this works: if assets are uncorrelated but have the same returns, then how does that lead to a better outcome but for re balancing?

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Tue Mar 14, 2017 3:23 pm

Dirghatamas wrote:I have a more serious/ operational question in terms of learning from you guys. How do you guys actually implement this in terms of accumulation or withdrawals? If there are lets say 2 assets in one's portfolio and they zig and zag and are uncorrelated, then on paper this doesn't help you, unless you re balance (move money between them) in accumulation or selectively take money out of them in withdrawal. However, the compelling data for re balancing between at least stocks and intermediate/short/total bonds is that it doesn't lead to better end value. There is some data on re balancing bonus between long bonds and stocks but long bonds are frowned on this forum (correctly so due to very high risk to interest rate).

So, if you combine say reasonable duration bonds and tilted stocks (say small cap value or whatever), aren't you simply relying on the assertion that the tilt will lead to higher returns, because small cap value will outperform? If not, are you guys re balancing between different stock classes?

This time I am not trying to argue, simply trying to understand how this works: if assets are uncorrelated but have the same returns, then how does that lead to a better outcome but for re balancing?


Hi Dir,

Thanks for the apology, that is kind of you. In this case no worries, as it's been a very interesting discussion, which I always enjoy.

Your question above gets to one of the reasons why I posted in the first place. When I was playing with a data file I'd found on the site (https://www.bogleheads.org/forum/viewtopic.php?t=93095), I started to worry about how much rebalancing I would need to do.

I've tried to answer your question with an example, by playing with the spreadsheet Joe mentioned. However, while playing with the data it raises a number of new questions for me -- but I'd hate to further derail my own thread. :)

To give a not completely satisfying example from that spreadsheet: in their model, 100% TSM gives you 5.9% return and a maximum portfolio drop of 51%, with 13 years to recover to your original amount. Small Caps give you a better return (6.9%), but with a 60% drop and 9 years to recover. Let's say that's too much risk for you. You can go to 80% SC and 20% bond. This gives 6.5% return, with only a 52% drop and 7 years to recover. So 0.6% more return than TSM with similar risk, and 0.4% worse return than SC, but with significantly better risk.

That says you can get a better fit to your risk / return with the SC + Bonds, but it also says you absolutely had better rebalance. Otherwise, without realizing it, you could be taking on far more risk than you are comfortable with as your portfolio slides towards a higher % SC. That is why I was so interested in the risk side of the equation, to get a feel for how careful I need to be with rebalancing if I tilt.

Sean

Dirghatamas
Posts: 506
Joined: Fri Jan 01, 2016 6:18 pm

Re: What do you expect to achieve with your Tilt?

Post by Dirghatamas » Tue Mar 14, 2017 4:05 pm

Just to add to my previous post, you need to additionally model two things

1) Tax drag from re balancing: Assuming taxable, the more volatile an asset e.g. small cap, the more you will need to re balance to keep risk within bounds. With 100% total stock portfolios, because there is no re balancing, there is no additional tax drag.

2) Behavioral aspect of re balancing: A high return/high volatility + low return/ low volatility only works if you DO re balance. 2008 is a good example of case studies. When stocks fall rapidly by say 50% or say 60% for even higher risk asset like SCV, the theory just assumes you will re balance. Very few people do. There are threads after threads on this forum on what people actually did. People DON'T re balance adequately back into stocks (usually not at all) when things look bleak. That is a very real aspect of investing that these theoretical studies usually miss.

100% stock portfolio (whether US or world or whatever) are brutal but simple. There is never an additional tax drag and (because you have no place to go) you can't make behavioral re balancing mistakes. During 2000-2002 and even more so in 2008, I was unhappy about the 50% draw down..but there wasn't anything to do, given my investing philosophy, so I couldn't make any behavioral mistakes.

PS: I do think the risks involved in a 80% SCV + 20% bonds are MUCH higher than a 100% TSM or even better a 100% World Stock.

User avatar
Robert T
Posts: 2432
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: What do you expect to achieve with your Tilt?

Post by Robert T » Tue Mar 14, 2017 4:33 pm

dbr wrote:But, Robert, what was your targeted result? The OP suggested at least the expected mean and expected SD, the usefulness of SD being a point for discussion. The point is that all results are only one instance from a possible distribution and we plan portfolios based on the expected distribution. The whole point of the question, it seems to me, is to specify in numbers what the rationale is for any person's adoption of a tilted portfolio and what that portfolio actually is. Actually would be what investments, and also what the factor loadings of those investments have been over the time concerned.

Long-term expectations (from start of 2003)

Annualized return = 7.5%
Standard deviation = 14
Loss in any one calendar year = 30-35%

Portfolio was constructed to align expected annualized return with needed return, and downside risk with tolerable loss.

When setting up my portfolio at start of 2003, I titled it to value and small caps, and to international to achieve my needed return. At the time, Bernstein’s expected returns for US stocks in his book Four Pillars of Investing = 6.5% (assuming a 3% inflation rate), so this was below my needed return. With a small cap and value tilt, together with foreign developed and emerging market stocks, and 25% bonds, I was able to align my portfolio expectations to both needed return and tolerable loss.

From the SD above, I should expect (or at least not be surprised by) negative returns every 6 years (1 SD), or at least a 20% loss every 44 years (2 SD). This assumes returns are normally distributed, but annual return distributions seem to have fatter tails so would not be surprised with more frequent losses.

The standard deviation also tells us more than just what to expect annually, it provides an indication of what to expect over longer time horizons e.g. 25 years. For example a SD of 14 implies returns of one in every six 25 year periods of 2.8 percent below the mean [14/SQRT(25)] i.e. there is about 16 percent odds that my portfolio return over a 25 year period may in fact be 4.7% (7.5 – 2.8). Compounded over time this is a huge difference. For each dollar invested at inception, the portfolio with a 4.7% return would accumulate into about half the value of a portfolio with a 7.5% return. This difference is reduced with additional annual savings over 25 years, but nevertheless a large difference. So I should not be surprised with this result if it materializes as this is what the standard deviation it telling me (just to note, while I mixed average and annualized returns, it still provides indicative outcomes).

FWIW I track actual progress relative to expected progress and the potential outcomes with a 16 percent odds of occurrence (implied by a standard deviation of 14). An example is provided in the chart in this earlier thread viewtopic.php?f=10&t=7353

Image

This, to me, illustrates the importance of also paying attention to SD to try to reduce the potential dispersion of portfolio outcomes (i.e. increase the odds of success) over an investment lifetime.

Obviously no guarantees, so also good to ensure other contingencies (e.g. perhaps when establishing needed return, initially plan on somewhat conservative estimates for savings, reasonable spending, and early retirement to allow scope to save more, spend less, work longer).

Robert
.

dbr
Posts: 22908
Joined: Sun Mar 04, 2007 9:50 am

Re: What do you expect to achieve with your Tilt?

Post by dbr » Tue Mar 14, 2017 4:52 pm

^^^ Now we are getting somewhere. Thank you for being specific. I also applaud your mention of the fact that short term volatility is in fact important to long term uncertainty (or volatility) because the long term is just the compounded result of the short term.

Thanks.

I shudder to refer to the other thread asking about designing portfolios nearer to the efficient frontier to gain better compound return from the same annual average return.

Theoretical
Posts: 1063
Joined: Tue Aug 19, 2014 10:09 pm

Re: What do you expect to achieve with your Tilt?

Post by Theoretical » Tue Mar 14, 2017 5:09 pm

In terms of risk profile, I treat 1 Larry Portfolio unit as 1.5 Equity (.6, .6 S/V) and 1 Robert T Portfolio unit (.2, .4 S/V) as 1.33 Equity, ergo my choice to be in the 60/40 range to target a risk profile in the 80/20 range.

So I agree that an 80/20 SCV/Bonds portfolio would be riskier - in fact, I'd treat it like a 120/20 portfolio.

User avatar
Wildebeest
Posts: 916
Joined: Fri Dec 27, 2013 2:36 pm

Re: What do you expect to achieve with your Tilt?

Post by Wildebeest » Tue Mar 14, 2017 6:39 pm

Robert T wrote:FWIW I track actual progress relative to expected progress and the potential outcomes with a 16 percent odds of occurrence (implied by a standard deviation of 14). An example is provided in the chart in this earlier thread viewtopic.php?f=10&t=7353

Image

This, to me, illustrates the importance of also paying attention to SD to try to reduce the potential dispersion of portfolio outcomes (i.e. increase the odds of success) over an investment lifetime.

Obviously no guarantees, so also good to ensure other contingencies (e.g. perhaps when establishing needed return, initially plan on somewhat conservative estimates for savings, reasonable spending, and early retirement to allow scope to save more, spend less, work longer).

Robert
.


Thanks RobertT for posting.

I scoured your earlier thread, which is one of my all time favorites as to your position on rebalancing. I read all the links again and I enjoyed reading them even while I had read them before and still much remained over my head.

I did not see any links regarding rebalancing. I assume you rebalance.

John C Bogle stated in The Bogle eBlog: Ask Jack June 15 2007

Excerpt

"We’ve just done a study for the NYTimes on rebalancing, so the subject is fresh in my mind. Fact: a 48%S&P 500, 16% small cap, 16% international, and 20% bond index, over the past 20 years, earned a 9.49% annual return without rebalancing and a 9.71% return if rebalanced annually. That’s worth describing as “noise,” and suggests that formulaic rebalancing with precision is not necessary.

We also did an earlier study of all 25-year periods beginning in 1826 (!), using a 50/50 US stock/bond portfolio, and found that annual rebalancing won in 52% of the 179 periods. Also, it seems to me, noise. Interestingly, failing to rebalance never cost more than about 50 basis points, but when that failure added return, the gains were often in the 200-300 basis point range; i.e., doing nothing has lost small but it has won big. (I’m asking my good right arm, Kevin, to send the detailed data to you.)

My personal conclusion. Rebalancing is a personal choice, not a choice that statistics can validate. There’s certainly nothing the matter with doing it (although I don’t do it myself), but also no reason to slavishly worry about small changes in the equity ratio. Maybe, for example, if your 50% equity position grew to, say, 55% or 60%.



Best,
Jack


I never re-balanced because it seemed to much work and is messy and I loved to read John C Bogle's quote and it so much easier to stay the course when I am oblivious to the market gyrations.

Do you re-balance ? What is your rationale? and if you do: what are your bands? Have you determined what your return would have been if you would not have re-balanced?

Thanks again for a great post viewtopic.php?f=10&t=7353.
The Golden Rule: One should treat others as one would like others to treat oneself.

User avatar
Portfolio7
Posts: 207
Joined: Tue Aug 02, 2016 3:53 am

Re: What do you expect to achieve with your Tilt?

Post by Portfolio7 » Wed Mar 15, 2017 2:37 am

Fwiw (I hope the formatting holds when I post!). The Simba spreadsheet seems to point to over 2 points of outperformance vs the three fund portfolio, for slightly less down side risk (if that's what down SD is telling me.) However, the last 30 years suggest maybe the spread has narrowed (?) to about 1 & 1/4 percent, give or take, with risk a little higher. Any outperformance may dissappear tomorrow, of course. I figure it's a bet with pretty good odds, though of course it still may fail. If it succeeds, the numbers look real nice. If not, there's still a good chance to match the 3 fund portfolio. I had to tweak the Simba spreadsheet a bit to reflect some of my investment options/history, so be aware the data is modified.

I'm still in mostly growth phase, but getting close enough to retirement that I'm starting to reduce risk.

Eventually my portfolio will, I think, reflect Larry Swedroe's recent columns somewhat, wherein the primary purpose will be to reduce fat tails... my portfolio analysis shows that when I adjust my portfolio for lower risk, the results also seem to reduce fat-tails, if I understand it all correctly. I assume it's because I'm targeting similar factors in similar fashion, at least to a modest degree. However, my portfolio below is designed to seek greater return relative to a 60/40 3 fund portfolio with roughly the same risk, so I would assert that in it's current form, my portfolio is significantly different from the low fat-tails portfolio.


1970-2015 1985-2015
3-Fund Portfolio7 3-Fund Portfolio7
Average 9.95% 12.22% 10.13% 11.51%
Std. Deviation 11.24% 12.01% 11.58% 12.46%
Down SD (vs MAR) 6.17% 6.03% 6.19% 6.40%
Up SD (vs MAR) 10.49% 12.53% 10.85% 12.32%
CAGR 9.36% 11.55% 9.51% 10.79%
Sharpe Ratio 0.20 0.38 0.33 0.42
Sortino Ratio 0.80 1.20 0.83 1.02
US Mkt. Correlation 0.94 0.83 0.93 0.82
Int'l Mkt. Correlation 0.82 0.89 0.86 0.90
Total - Rebalanced (N) 612,932 1,524,329 167,153 239,605
Total-Unbalanced (N) 549,034 1,677,818 159,325 238,540
Total - Rebalanced (Real) 97,696 242,965 74,416 106,671
An investment in knowledge pays the best interest.

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Fri Mar 17, 2017 8:23 am

I guess I shouldn't travel when I have an open topic -- I get behind!

In any case, it looks like the thread as run its course. Thanks for all the helpful inputs. A couple of replies to the last posts.

Dirghatamas wrote:Just to add to my previous post, you need to additionally model two things

1) Tax drag from re balancing

2) Behavioral aspect of re balancing

Yes, emphasizing the point that you need to rebalance

PS: I do think the risks involved in a 80% SCV + 20% bonds are MUCH higher than a 100% TSM or even better a 100% World Stock.

I just used the link Joe gave us, and don't have an opinion on this. It underlines my belief that if you don't at least try to quantify expected return and risk, you can't really have a discussion or make choices


Robert T wrote:Long-term expectations (from start of 2003)

Annualized return = 7.5%
Standard deviation = 14
Loss in any one calendar year = 30-35%


FWIW I track actual progress relative to expected progress and the potential outcomes with a 16 percent odds of occurrence (implied by a standard deviation of 14). An example is provided in the chart in this earlier thread viewtopic.php?f=10&t=7353

Robert
.


This was a really useful post. Thanks, Robert.

Wildebeest wrote:
Robert T wrote:
John C Bogle stated in The Bogle eBlog: Ask Jack June 15 2007

I never re-balanced because it seemed to much work and is messy and I loved to read John C Bogle's quote and it so much easier to stay the course when I am oblivious to the market gyrations.



Thanks for the post, Beest. It is very interesting, and not in line with my intuition. I guess since you are talking "percents of percents" it really is a second order effect compared with AA. to be honest, rebalancing does not seem that difficult to me: if I am accumulating, I just put in a bit more or less per fund; if withdrawing, take more or less out. Of course it is relatively easy with just three funds. :-)

User avatar
Robert T
Posts: 2432
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Re: What do you expect to achieve with your Tilt?

Post by Robert T » Fri Mar 17, 2017 9:05 pm

Wildebeest wrote:Do you re-balance ? What is your rationale? and if you do: what are your bands? Have you determined what your return would have been if you would not have re-balanced?

Wildebeest,

Yes I rebalance, to rebalance risk.

I have long-term targets for allocations to stocks:bonds, US:EAFE:EM; and exposure to value, size and term – based on willingness, ability, and need to take risk. So I try to stick with this over the long-term (why have targets if don’t stick to them)?

If I had not rebalanced since start in 2003 (vs. rebalancing) it would have resulted in 0.3% lower return, larger 2008 draw down and higher volatility.

Note Balanced / Rebalanced

..9.0% / ..9.3% = Annualized return
-34.8 / -28.7 = 2008 downside
..17.6 / ..15.9 Standard deviation

My stock:bond allocation would now be 88:12 compared to my 75:25 long term target.

The only thing Bogle talks about in the quote you referenced is annualized returns over 20 to 25 year periods, nothing about volatility or drawdowns. If not concerned about the latter two then why not simply hold 100% small value stocks for highest expected return (as the only metric).

Personally I am not only concerned about expected returns, but also about annual drawdowns (to help stay the course), and volatility (expected dispersion of long-term returns). Rebalancing helps keep these latter two risks in-line with risk tolerance.

There are a myriad of ways to rebalance, some more complicated than others:

    - Calendar year rebalancing (annual, or every two to three years [re; Bernstein])
    - 5x25 bands [re: Larry Swedroe],
    - Adjusting bands by asset class to reflect the distribution of returns for each asset class (using wider bands for more volatile asset classes), with bands on the upside being larger than the downside [re; Bill Bernstein],
    - Valuation based rebalancing (overweighting undervalued asset classes, but within bands),
    - Momentum based rebalancing (overweighting asset classes with one year absolute and relative momentum (re: dual momentum), but within bands,
    - And perhaps even a combination of valuation and momentum based.
As Bernstein says: “Rebalancing requires nerves and discipline; overbalancing requires even more of both of these scarce commodities. Very few investors, small or institutional, can carry it off”.

I think simple annual rebalancing is fairly tough to beat (easier to stick with over long term). Obviously choice of rebalancing depends on the share of a portfolio in tax and tax-advantaged accounts. In any event, directing new savings to underweighted asset classes can help keep these more aligned to long-term targets.

Robert
.

columbia
Posts: 474
Joined: Tue Aug 27, 2013 5:30 am

Re: What do you expect to achieve with your Tilt?

Post by columbia » Fri Mar 17, 2017 9:17 pm

It seems that the universal answer should be: greater volatility and returns.

User avatar
Wildebeest
Posts: 916
Joined: Fri Dec 27, 2013 2:36 pm

Re: What do you expect to achieve with your Tilt?

Post by Wildebeest » Fri Mar 17, 2017 10:06 pm

Robert T wrote:
Wildebeest wrote:Do you re-balance ? What is your rationale? and if you do: what are your bands? Have you determined what your return would have been if you would not have re-balanced?

Wildebeest,

Yes I rebalance, to rebalance risk.

I have long-term targets for allocations to stocks:bonds, US:EAFE:EM; and exposure to value, size and term – based on willingness, ability, and need to take risk. So I try to stick with this over the long-term (why have targets if don’t stick to them)?

If I had not rebalanced since start in 2003 (vs. rebalancing) it would have resulted in 0.3% lower return, larger 2008 draw down and higher volatility.

Note Balanced / Rebalanced

..9.0% / ..9.3% = Annualized return
-34.8 / -28.7 = 2008 downside
..17.6 / ..15.9 Standard deviation

My stock:bond allocation would now be 88:12 compared to my 75:25 long term target.

The only thing Bogle talks about in the quote you referenced is annualized returns over 20 to 25 year periods, nothing about volatility or drawdowns. If not concerned about the latter two then why not simply hold 100% small value stocks for highest expected return (as the only metric).

Personally I am not only concerned about expected returns, but also about annual drawdowns (to help stay the course), and volatility (expected dispersion of long-term returns). Rebalancing helps keep these latter two risks in-line with risk tolerance.

There are a myriad of ways to rebalance, some more complicated than others:

    - Calendar year rebalancing (annual, or every two to three years [re; Bernstein])
    - 5x25 bands [re: Larry Swedroe],
    - Adjusting bands by asset class to reflect the distribution of returns for each asset class (using wider bands for more volatile asset classes), with bands on the upside being larger than the downside [re; Bill Bernstein],
    - Valuation based rebalancing (overweighting undervalued asset classes, but within bands),
    - Momentum based rebalancing (overweighting asset classes with one year absolute and relative momentum (re: dual momentum), but within bands,
    - And perhaps even a combination of valuation and momentum based.
As Bernstein says: “Rebalancing requires nerves and discipline; overbalancing requires even more of both of these scarce commodities. Very few investors, small or institutional, can carry it off”.

I think simple annual rebalancing is fairly tough to beat (easier to stick with over long term). Obviously choice of rebalancing depends on the share of a portfolio in tax and tax-advantaged accounts. In any event, directing new savings to underweighted asset classes can help keep these more aligned to long-term targets.

Robert
.


Robert T,

I really appreciate the rigor of your data analysis. I had hoped your data would show that it did not make any difference and my Laissez faire approach of: "Set it and forget it", would do me well.

If I had the where it all, I would be able to figure out if not have having bonds and being significantly tilted to small value and emerging markets ( more so in the tax advantaged accounts) would to do as to comparison of re-balancing versus "the set and forget it" for the last 10 years. But it still may not have answer the question as what will be better for the next 10 or 20 years.

Your data is convincing for the last 14 years: The 0.3% higher return on re-balancing in a 75 equities /25 % bonds makes me wonder if I should not just roll the dice, but instead embrace rebalancing.

Again thanks for such an excellent response to my question.

Wildebeest.
The Golden Rule: One should treat others as one would like others to treat oneself.

User avatar
Sammy_M
Posts: 1855
Joined: Sun Nov 25, 2007 8:30 am

Re: What do you expect to achieve with your Tilt?

Post by Sammy_M » Mon Mar 20, 2017 6:36 pm

Robert T wrote:The standard deviation also tells us more than just what to expect annually, it provides an indication of what to expect over longer time horizons e.g. 25 years. For example a SD of 14 implies returns of one in every six 25 year periods of 2.8 percent below the mean [14/SQRT(25)] i.e. there is about 16 percent odds that my portfolio return over a 25 year period may in fact be 4.7% (7.5 – 2.8). Compounded over time this is a huge difference. For each dollar invested at inception, the portfolio with a 4.7% return would accumulate into about half the value of a portfolio with a 7.5% return. This difference is reduced with additional annual savings over 25 years, but nevertheless a large difference. So I should not be surprised with this result if it materializes as this is what the standard deviation it telling me (just to note, while I mixed average and annualized returns, it still provides indicative outcomes).


dbr wrote:^^^ Now we are getting somewhere. Thank you for being specific. I also applaud your mention of the fact that short term volatility is in fact important to long term uncertainty (or volatility) because the long term is just the compounded result of the short term.


I find this very interesting and would like to understand it better. Would you please elaborate or point me to any further posts on the subject or educational materials? Thank you!

dbr
Posts: 22908
Joined: Sun Mar 04, 2007 9:50 am

Re: What do you expect to achieve with your Tilt?

Post by dbr » Tue Mar 21, 2017 8:27 am

Sammy_M wrote:
Robert T wrote:The standard deviation also tells us more than just what to expect annually, it provides an indication of what to expect over longer time horizons e.g. 25 years. For example a SD of 14 implies returns of one in every six 25 year periods of 2.8 percent below the mean [14/SQRT(25)] i.e. there is about 16 percent odds that my portfolio return over a 25 year period may in fact be 4.7% (7.5 – 2.8). Compounded over time this is a huge difference. For each dollar invested at inception, the portfolio with a 4.7% return would accumulate into about half the value of a portfolio with a 7.5% return. This difference is reduced with additional annual savings over 25 years, but nevertheless a large difference. So I should not be surprised with this result if it materializes as this is what the standard deviation it telling me (just to note, while I mixed average and annualized returns, it still provides indicative outcomes).


dbr wrote:^^^ Now we are getting somewhere. Thank you for being specific. I also applaud your mention of the fact that short term volatility is in fact important to long term uncertainty (or volatility) because the long term is just the compounded result of the short term.


I find this very interesting and would like to understand it better. Would you please elaborate or point me to any further posts on the subject or educational materials? Thank you!


A good illustration is here: http://www.norstad.org/finance/risk-and ... l#appendix

The comment is often made that this illustration based on a random walk exaggerates the long term variability because actual investment results have return to the mean behavior. While it is true the variability is reduced due to this, the variability still grows but not as fast as sqrt(time). Also, the illustration is for a normal distribution. Actual investments have wilder than normal distribution behavior. In addition, the example assumes the same distribution of annual returns while the mean and variability of annual returns may itself vary over time.

A good way to see the effect in real historical results is to go into FireCalc, enter a period of time with no withdrawals, or with them for that matter, and just look at the range of outcomes for the value of the portfolio as time goes by. You can play with the portfolio going from 100% S&P 500 to 100% 3 mo. T bills.

IlliniDave
Posts: 2117
Joined: Fri May 17, 2013 7:09 am

Re: What do you expect to achieve with your Tilt?

Post by IlliniDave » Tue Mar 21, 2017 8:46 am

So, I didn't read all the prior posts.

The main reason for my tilt is just to sate the part of me that feels like it needs to do something, without actually doing much of anything (once a tilt is chosen it's managed passively). I hope to get maybe an extra 0.5% return over the balance of my lifetime for the tilt (a little small-valuey and a little towards EM on the overseas side). Not really sophisticated or aggressive. Getting a little extra return may or may not work out.
Don't do something. Just stand there!

lgs88
Posts: 206
Joined: Tue Jun 07, 2016 7:48 am

Re: What do you expect to achieve with your Tilt?

Post by lgs88 » Tue Mar 21, 2017 8:47 am

I expect to achieve tracking error :D :D :D
merely an interested amateur

dbr
Posts: 22908
Joined: Sun Mar 04, 2007 9:50 am

Re: What do you expect to achieve with your Tilt?

Post by dbr » Tue Mar 21, 2017 9:14 am

If all people are looking for is a little more return while paying no attention to risk, then why not allocate a little more to stocks and a little less to bonds. If one is already 100% stocks that is a different issue.

Data supplied looking at both risk and return is on point, of course.

Random Walker
Posts: 1923
Joined: Fri Feb 23, 2007 8:21 pm

Re: What do you expect to achieve with your Tilt?

Post by Random Walker » Tue Mar 21, 2017 10:13 am

Dbr,
The reason to tilt instead of increase stock exposure when overall equity allocation is less than 100% is to diversify sources of risk. Increasing equity allocation just adds more of the same risk, market risk. Adding small and value, adds unique risks that are weakly correlated with market risk. The portfolio is more efficient. The tradeoff of course is cost. Market beta exposure is cheapest. Slightly more expensive to tilt to small and value.

Dave

dbr
Posts: 22908
Joined: Sun Mar 04, 2007 9:50 am

Re: What do you expect to achieve with your Tilt?

Post by dbr » Tue Mar 21, 2017 2:51 pm

Random Walker wrote:Dbr,
The reason to tilt instead of increase stock exposure when overall equity allocation is less than 100% is to diversify sources of risk. Increasing equity allocation just adds more of the same risk, market risk. Adding small and value, adds unique risks that are weakly correlated with market risk. The portfolio is more efficient. The tradeoff of course is cost. Market beta exposure is cheapest. Slightly more expensive to tilt to small and value.

Dave


Yes, I know that as it is stated. What is missing and is to the point of the OP on this thread is to demonstrate that in actual cases showing numbers. The numbers in question would be the distribution of returns expected year after year. People cannot continually present to terminology such as "diversify" and "sources of risk" without eventually producing the only real item in the whole discussion which is what return materializes year by year, or, more directly, what is the value of the portfolio over time, prospectively.

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Thu Mar 30, 2017 6:38 am

So thanks all for the replies. While it was disappointing to get so few concrete answers to my OP, I've gotten a lot of what I was after. My tentative top level conclusion is that I'm going to shift my portfolio from three funds to something like:

85/15 Stock/Bond
50/50 Total market/Mid&SmValue
< 30 USD (I live in Europe, and have had dramatic experiences with currency fluctuations. I take them seriously)

Possibly about five funds at that point. I need to dig a bit into how to implement and calculate my expected numbers, as < 30% US limits data availability. If anyone has a good link to a (very) simple how-to on the mechanics of using Fama-French data in a spreadsheet, I would certainly be grateful. The ones I have found so far don't get down to the "copy file xyz into cell B3" level that I'm hoping for.



Some side topics:

AlohaJoe wrote:Sure, I try to quantify it. Or, maybe "quantify" isn't quite the right word because it implies boiling it down to single, simple numbers. But I tend to prefer comparisons of drawdowns & income equivalents.

I've started looking at seven numbers as I compare portfolios: CAGR 70 - 16, 87 - 16, 07 - 08 and 07 - 16; Max draw-down; recovery time; standard deviation. I still like SD, although am happy having Joe's extra parameters: SD is easy to calculate and compare and I'm still convinced that using the delta SD to compare portfolios nets out a pretty large chunk of the various non-normal problems.


Robert T wrote:An example is provided in the chart in this earlier thread viewtopic.php?f=10&t=7353

This is an incredible post. I assume that once I've plowed through it all, there is some sort of a ceremony and I get a signed certificate. :happy


dbr wrote:If all people are looking for is a little more return while paying no attention to risk, then why not allocate a little more to stocks and a little less to bonds.

+1 I just don't understand deviating from TSM without some quantifiable expectation of an improvement to risk/reward.

cheers,
Sean
Last edited by sean.mcgrath on Thu Mar 30, 2017 6:42 am, edited 1 time in total.

User avatar
happymob
Posts: 536
Joined: Wed Nov 18, 2009 4:09 pm

Re: What do you expect to achieve with your Tilt?

Post by happymob » Thu Mar 30, 2017 6:41 am

Honestly, I may or may not achieve any actual improvement in returns.

What I expect to achieve is a little added entertainment as I see that small-cap value is up (or down) or emerging markets are up (or down). It's just more interesting than a 3-fund portfolio.

User avatar
willthrill81
Posts: 2210
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: What do you expect to achieve with your Tilt?

Post by willthrill81 » Thu Mar 30, 2017 9:05 am

sean.mcgrath wrote:
dbr wrote:If all people are looking for is a little more return while paying no attention to risk, then why not allocate a little more to stocks and a little less to bonds.

+1 I just don't understand deviating from TSM without some quantifiable expectation of an improvement to risk/reward.


I'm not sure why you want a "quantifiable expectation of an improvement" when we don't even know what the returns of the TSM (assumed 'baseline') will be going forward.

You seem to be wanting a prediction of the future.

Part of the (at least historical) premiums from factors like small cap comes from the uncertainty of what the premium will be going forward. And the premium seems to be constantly shifting, which averages neither capture nor illustrate.

For instance, from 1972-1979, the nominal premium for SC over TSM was around 5%. From 1980 to 1989, the premium was around .6%. From 1990 to 1999, the premium was a negative 3%. From 2000-2009, the premium was about 4.6%. And from 2010 to current, the premium has been about .7%.

My interpretation of those data is that, at least SC, is likely to continue to have a premium going forward over a sufficiently long (+10 years) time frame, but it's likely impossible to know how much that premium will be.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Thu Mar 30, 2017 10:01 am

willthrill81 wrote:I'm not sure why you want a "quantifiable expectation of an improvement" when we don't even know what the returns of the TSM (assumed 'baseline') will be going forward.

You seem to be wanting a prediction of the future.


For me "expected" in probability theory is the median or average result of a variable in a model. Of course no one expects to actually achieve their expected values -- models are simply a way of thinking about a problem. Unless we simply follow a guru without thinking, we all have models based on history that we believe give us some understanding about likely future outcomes.

My point is that, if I decide to S&D instead of simply follow TSM, I need a justification based on history and perhaps theory for my deviation, and the justification needs to include both risk and return. Otherwise I'm simply following a random guru. It doesn't need to be a complicated model, but it does need to be coherent.

User avatar
willthrill81
Posts: 2210
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: What do you expect to achieve with your Tilt?

Post by willthrill81 » Thu Mar 30, 2017 11:35 am

sean.mcgrath wrote:
willthrill81 wrote:I'm not sure why you want a "quantifiable expectation of an improvement" when we don't even know what the returns of the TSM (assumed 'baseline') will be going forward.

You seem to be wanting a prediction of the future.


For me "expected" in probability theory is the median or average result of a variable in a model. Of course no one expects to actually achieve their expected values -- models are simply a way of thinking about a problem. Unless we simply follow a guru without thinking, we all have models based on history that we believe give us some understanding about likely future outcomes.

My point is that, if I decide to S&D instead of simply follow TSM, I need a justification based on history and perhaps theory for my deviation, and the justification needs to include both risk and return. Otherwise I'm simply following a random guru. It doesn't need to be a complicated model, but it does need to be coherent.


Small cap and value factors both have historical evidence as well as theoretical justification going for them. Are you looking for a mathematical model?
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

sean.mcgrath
Posts: 104
Joined: Thu Dec 29, 2016 6:15 am
Location: US in NL

Re: What do you expect to achieve with your Tilt?

Post by sean.mcgrath » Thu Mar 30, 2017 2:37 pm

willthrill81 wrote:Small cap and value factors both have historical evidence as well as theoretical justification going for them. Are you looking for a mathematical model?


I'm not quite sure where we have a disconnect. My purpose of the OP was, since was I thinking about moving from a total market approach to S&D, to get a picture of what other people expected to gain from their S&D approach as I was debating whether it was worth my time to implement it. I'm pretty flexible about what constitutes a model, and was expecting answers ranging from rough ideas like "a bit more return with no additional risk" to spreadsheet results like "+1.5% return with -0.2% SD." I was surprised that so few people would articulate the risk part of their expectations.

I don't believe you can think coherently about investment return if you don't think about the associated risk. I buy into the small cap and value theories, but I personally would not switch my investment style without having a rough idea of the rationale and the expected trade-offs. And yes, as I posted, I do plan to move to a tilt :-)

dbr
Posts: 22908
Joined: Sun Mar 04, 2007 9:50 am

Re: What do you expect to achieve with your Tilt?

Post by dbr » Thu Mar 30, 2017 3:37 pm

willthrill81 wrote:
Small cap and value factors both have historical evidence as well as theoretical justification going for them. Are you looking for a mathematical model?


Of course there are models for the average annual return of various portfolios as a function of factor loadings. I am not sure I have seen a model for the annual standard deviation of returns for various portfolios as a function of factor loadings. Maybe I have overlooked somewhere that this is presented.

Post Reply