Portfolio Charts

[Tyler9000]

Yeah, I suppose it's also a good time to dig back through the suggestion box. There are several good ideas still floating around.

January should definitely be busy with the new 2015 data and the various updates that entails. I hope to start adding a few more new things as well.

[jennypenny]

I like the new section explaining the asset classes.

-------------

I probably just need more coffee, but I can't figure out why the sustainable WR for this chart is so much lower for a 10yr duration. What am I missing?

[FBeyer]

The one thing I'd love to have, is the ability to type in one asset allocation, and then get output from the entire range of figures one can produce.
Like FIretime, pixels, and safeWR chart all at once, say.

[fips]

I like the new section explaining the asses classes.

I probably just need more coffee, but I can't figure out why the sustainable WR for this chart is so much lower for a 10yr duration. What am I missing?

Methodically, the sustainable WR shows the worst possible time range, as explained here.

[jennypenny]

I like the new section explaining the asses classes.

Yeah, I really need more coffee this morning.

[Tyler9000]

I like the new section explaining the asset classes.

-------------

I probably just need more coffee, but I can't figure out why the sustainable WR for this chart is so much lower for a 10yr duration. What am I missing?

Thanks!

The Sustainable WR is the one that maintained the initial inflation-adjusted principal at the end of the single worst stated timeframe. If you plug the same AA into the Funnel chart, you can see how SustWRs kinda-sorta track the red line (it's more complicated than that, but you get the idea).

Looking behind the calculation curtain, the single worst 10-year SustWR period for that portfolio started in 1972. It was just a rough period for that combination of assets (many combinations, actually -- lots of SustWRs are negative over that short of a timeframe).

Give them a few more years to average out, and results pick up quickly. But FWIW I do like to use that 10-year SustWR as a decent measure of whether someone will actually stick with the plan long enough for that to happen. Most people can't handle losing money for a decade without changing course and starting over with something new.

[Tyler9000]

The one thing I'd love to have, is the ability to type in one asset allocation, and then get output from the entire range of figures one can produce.

Yeah, I made a spreadsheet for myself that does just that to help generate the summary charts. It's just a tricky balance between ease-of-use for entering everything only once and ease-of-comprehension by presenting everything separately. The fact that some of the heavier calculations slow everything down also doesn't help.

I do find myself missing the "CAGR since 1972" number on the pixel chart and having to hover my mouse on the right cell or flip over to the Funnel, so I might add that back. What similar pieces of data are you commonly wanting to see together?

BTW, one shortcut you can use is to copy/paste the entire column of black cells between calculators. That's one of the perks of using an Excel app.

[fips]

What kinds of asset allocation problems would you like to be able to solve?

Since you asked for it ...

Here is an easy and specific one to start with:
What portfolio of three evenly distributed assets provide the highest overall CAGR since 1972 (regardless of volatility)?

[arebelspy]

What kinds of asset allocation problems would you like to be able to solve?

Since you asked for it ...

Here is an easy and specific one to start with:
What portfolio of three evenly distributed assets provide the highest overall CAGR since 1972 (regardless of volatility)?

A "seek" feature like this would be neat.

I put in a minimum CAGR I'm satisfied with, for example, and search for the portfolio(s) that hits that, with the lowest volatility.

Or vice-versa (put how much volatility I'm fine with, and search the portfolio that gives the highest CAGR).

Bonus points if I can uncheck assets, so it only searches the ones I want.

I imagine the computing power (and thus time it'd take to run this) might be prohibitive though.

[jacob]

I put in a minimum CAGR I'm satisfied with, for example, and search for the portfolio(s) that hits that, with the lowest volatility.

Or vice-versa (put how much volatility I'm fine with, and search the portfolio that gives the highest CAGR).

If this is possible, it would be more interesting/practical to look at Treynor or even better Sortino ratios. They're more practical than volatility from the individual investor's perspective. Mutual funds may worry about volatility since they have to manage value flows to both the upside and the downside. The Sortino ratio only considers losses to be penalizing when evaluating performance.---Most investors (as opposed to money managers) are quite happy when their portfolio swings to the upside even if that increases volatility.

[Tyler9000]

@ Arebelspy -- I've been playing with a spreadsheet that does exactly that (including letting you exclude assets). But for a few reasons I don't yet consider it ready to publish.

1) The results get a little fussy sometimes, as because of the number of variables I suspect there are multiple similar solutions from a convergence perspective. So you don't always get the same result every time.

2) The brute force method requires a solver plugin that won't work in the web tools I use. After experimenting a bit, I also don't personally care for the classic efficient frontier calculations as not all portfolio returns are best described by randomly generated normal distributions. I really like using actual historical sequence of returns.

3) The "optimized" portfolios aren't necessarily options I'd recommend. The thing I like about the Portfolios section is that I don't think any of them are bad options (or I wouldn't offer them). Since I can't necessarily say the same about non-curated optimizer output, I'm trying to find a good way to message the results before tacitly endorsing them. That could be either a really good visualization that captures the risk or a different angle to the original question.

I think I've found a good workaround for 1 and 2 with a new solving method. The Assets section is sorta my first step to address problem #3, but I'm still working on that part.

@ Jacob: Sortino ratios are interesting and I agree that they are valuable, but the trick is describing them in a way that makes sense to the average investor. (One goal of the latest Frequency chart is to visualize standard deviation -- another valuable but esoteric number for most people). Maybe a similar thing I can try is to find the combination of assets with the maximum return for a user-selected maximum drawdown.

@ Fips: I really like that question, although I may have to re-think my algorithm. A similar question I've been asking myself is "given two holdings already in a portfolio, what third asset would you recommend to most improve performance?"

TL;DR -- I'm working on it!

[JamesR]

Hey Tyler9000,

Portfoliocharts is great, although I found it a bit disconcerting at the complete lack of personal identity which makes it come off almost corporate-y. (have you see the cluetrain manifesto?) If it helps, you could just create a fake mascot/identify/whatever.. Maybe a logo of a Tyler9000 bot.

One thing I've noticed is that you are worried about the average person changing between portfolios, i.e. it could be expensive to do so. Perhaps there's ways around that, for someone that's still saving, they could target a new portfolio while maintaining aspects of the old portfolio that fit into the new one. Could be a tool there, given source portfolio(s) how to get to destination portfolio/allocation. Post-FI people might just be hooped?

After seeing the golden butterfly article, I'm tempted play around with portfolio allocations using something like genetic algorithms. Are you willing to share how you calculate the WR for the different allocations?

[Tyler9000]

Hmm... I'll have to think about the personal identity feedback. The site is intentionally not about me but about sharing a new way of looking at asset allocation. The goal is to build something useful that people will bookmark as a regular resource rather than create a typical blog that only lasts as long as you keep posting new content. People truly interested in the guy behind the numbers can generally track me down here or elsewhere, and I'm happy to share. I hear where you're coming from, though, and will have to think about how to express the right balance between professional and authentic.

Someone else has also suggested a glide path calculator for transitioning between portfolios. I'm still trying to figure out what that will look like and what output is most valuable, but I like the idea.

You can find the basic info on how the WRs are calculated here. When it comes to the specific calculations, the most succinct explanation is that I calculate the inflation-adjusted retirement growth (including withdrawals) beginning in every successive start year (44 separate retirements), create an iterative loop that adjusts each WR until the final balance reaches the desired value (0 for the SafeWR, initial principal for the SustWR), and find the lowest WR of each retirement length. It's the same basic method as the old Bengen SAFEMAX study, only automated. There are a few nuances, of course, but that should get you started.

[stayhigh]

Just wondering.

1. In theory, when I have a typical 80/20 portfolio, I should withdraw each year 3.2% stocks/0.8% bonds to get 4% WR, no matter how both of them perform. What happen, if I decide to withdraw better performing asset. During stock market prosperity, I will take 4% of stocks each year, without touching bonds. During crash, I will leave diving stocks and withdraw only bonds. In theory, that should give me about 5 years protection for my stocks during massive crash, before I run out of bonds. After that, I slowly will rebuild my 20% bonds allocation, by selling expensive stocks when market recover. Is that a solution for deviation and not selling valuable assets when they are cheap, or I will get nothing?

It also can be done for more complex portfolios, instead of automatic rebalancing to keep ratios nice, I would just sell few best performing assets and keep underperformers.

2. How about swapping Large Caps for Dogs of the Dow? Will it boost CAGR a little bit, or it's just unnecessary for long run?

[JamesR]

I just noticed your WR calculator update wasn't included with your FIREtime calculator apparently? For instance, when I try the 40% SCV, 40% EM, 20% GOLD allocation, it says "8.8% Sust WR" instead of the 7.3% the WR calculator shows.

[Tyler9000]

I just noticed your WR calculator update wasn't included with your FIREtime calculator apparently? For instance, when I try the 40% SCV, 40% EM, 20% GOLD allocation, it says "8.8% Sust WR" instead of the 7.3% the WR calculator shows.

Excellent catch! Apparently I neglected one part of the new logic when I updated the calculator for the new year. It should now work as intended.

Thanks for finding that and pointing it out.

[Tyler9000]

Just wondering.

1. In theory, when I have a typical 80/20 portfolio, I should withdraw each year 3.2% stocks/0.8% bonds to get 4% WR, no matter how both of them perform. What happen, if I decide to withdraw better performing asset. ?

From a calculation perspective, the working assumption is that you rebalance your portfolio and withdraw your expenses once a year at the same time. Practically speaking, that means that it generally makes the most sense to fund your expenses from the better performing asset first. Unless you're tax-loss harvesting, why sell bonds low and create a taxable event just to immediately buy them back in the rebalance? Likewise, in accumulation I usually recommend to apply new money to the lowest performing asset to rebalance on-the-fly and avoid unnecessary capital gains taxes.

I don't have enough data on Dogs of the Dow to have an opinion. If you know a good source with data since 1972, please drop me a PM.

[Tyler9000]

Here is an easy and specific one to start with:
What portfolio of three evenly distributed assets provide the highest overall CAGR since 1972 (regardless of volatility)?

Drumroll.....

The answer is small cap value, emerging markets, and international small. Real CAGR of 10.33% since 1972, although the volatility is very high and it once lost over 43% in a single year. It's not for your average investor.

BTW, figuring out how to answer this question destroyed the logjam and resulted in a ridiculous flood of interesting data. For example, the best performing portfolio of up to five equally weighted assets with no annual loss greater than 10% was international small, 5-year treasuries, short term treasuries, and gold (5.82% CAGR). And none of the top 800 options with the 10% worst year filter had less than four assets. Modern portfolio theory FTW.

The next challenge is how to present all of this. Thanks for the inspiration!

[fips]

The answer is small cap value, emerging markets, and international small. Real CAGR of 10.33% since 1972, although the volatility is very high and it once lost over 43% in a single year. It's not for your average investor.

Thanks for getting back to that! More evidence for correlation between risk and return.

The next challenge is how to present all of this. Thanks for the inspiration!

I think you could take portfoliocharts.com to the next level by introducing problem-solving charts that revolve around minimizing volatility, maximizing return or combinations thereof based on the investor's personal preference.

Keep up the good work!

[ThisDinosaur]

Wow, I'd love to see the list of "top 800 options" with the 10% filter. Any particular reason not to endorse the "Drumroll" portfolio? You've said previously, I think, that a portfolio derived this way, exclusively through back test optimization, wouldn't be recommended. And your Assets section on EM points out that they haven't been doing well in the past two decades. I'm curious how you, personally, use this information going forward, Tyler. Do you plan on changing your asset allocation somehow once you've finished tinkering with algorithms?

Its also interesting how similar that Best Performing Portfolio is to the PP and totally excludes the US market!