Early Retirement Extreme Forums

I just read this article which is pretty close to what I'm getting at:- http://monevator.com/is-any-asset-allocation-best/

The key point is this comment:- "But remember two things.

Firstly, hindsight and survivorship bias both loom large – the lop-sided asset allocations we remember are the ones that worked, not the ones that went nowhere.

Secondly – again – who knows which strongly differentiated allocation will win over the next 40 years?"

@steveo73: the reason brute is so interested and aggressive in talking about this, yet careful and questioning, is that he has the feeling either he is missing something, or many humans are missing something.

brute completely agrees that past performance doesn't guarantee future performance. but neither does anything else. so the common solution is to diversify. but since PPDGFP, what to diversify in? "the market" sounds cool but is not an answer unless a specific definition of "the market" is also given. domestic stock? domestic stock and bonds? GMP? what do humans base this definition on if not past performance? most indexers brute sees arguing actually do use past performance and then come up with "100% - how much of a coward a human is" domestic stock/bond.

so isn't the conclusion that all humans will be disappointed, because nobody had a crystal ball? the only human who will have had success will have gotten lucky. even the humans that "bought the market", they just got lucky with their definition of "the market".

is brute missing anything?

steveo73 wrote:I just read this article which is pretty close to what I'm getting at:- http://monevator.com/is-any-asset-allocation-best/

The key point is this comment:- "But remember two things.

Firstly, hindsight and survivorship bias both loom large – the lop-sided asset allocations we remember are the ones that worked, not the ones that went nowhere.

Secondly – again – who knows which strongly differentiated allocation will win over the next 40 years?"

I think we're mostly in the same page on what constitutes a good portfolio. We're down how to realistically compare them, and I believe the disconnect is a simple matter of micro vs macro perspectives. I agree that returns are unpredictable and that cherry picking individual historical timeframes to try to predict the top performer in the future is a really bad idea. However, I argue that since not every portfolio has the same uncertainty one can use history to make good (if not precise) decisions.

For example, consider these two different retirement portfolios. These charts show the historical portfolio values for a retiree withdrawing the traditional 4% a year, and they overlay 44 different retirement paths starting in every year since 1972. I've removed the portfolio names to separate the evidence from any particular dogma.




If one were to pick a single year to compare these two portfolios and isolate one line out of the mess, it's clear that the "best" portfolio is highly unpredictable. Option 1 may look much better than Option 2 one year, and the next year they may switch. And if one were to try to predict with certainty which one will have the better returns over the next 30 years, that would be a similarly futile exercise.

However when you look at all of the returns as a whole, a pattern emerges. All asset allocations are inherently unpredictable, but some are less so than others. By understanding the nuances of uncertainty rather than running from it, one can select a portfolio that clearly has a higher probability of achieving your goals. The differences in average and standard deviation (spread) are inherent qualities of a particular portfolio and are not simple blind luck. Asset allocation matters.

"Best" means different things to different people. I agree that those who want the single best return over their investing lifetime will not find much help in backtesting. But those who want the best probability of success for their particular life goals regardless of the variable return in any individual year along the way can learn a lot from history. The utility of the data depends heavily on what you're trying to get out of it.

Tyler9000 wrote: However when you look at all of the returns as a whole, a pattern emerges. All asset allocations are inherently unpredictable, but some are less so than others. By understanding the nuances of uncertainty rather than running from it, one can select a portfolio that clearly has a higher probability of achieving your goals. The differences in average and standard deviation (spread) are inherent qualities of a particular portfolio and are not simple blind luck. Asset allocation matters.

Yes that (my bolding)... but also to repeat this insight when one substitutes "asset allocation" methods with "stock picking" methods: Here's Buffett http://www8.gsb.columbia.edu/rtfiles/cb ... tt1984.pdf ... using pretty much the same argument you made but for stocks instead of asset classes. Like "how all stock picks are inherently unpredictable but some are less so than others". ... "That it's impossible to pick the one best portfolio but that specific classes of picking have differences in average and sds according to the inherent quality of the investment method and thus aren't simply blind luck".

It's crazy predictable how in these discussions, there's a strong tradition for running from disagreeable concepts by declaring them to be luck rather than understanding them or just acknowledging that other people might be capable of such understanding. That is to say any given investor is willing to agree that there's more to it than just luck exactly up to the point/level of nuance they personally understand ... beyond that point, it's all luck as far as they and everybody else is concerned. Much like resolving a fractal and declaring any detail beyond one's personal limit of resolution to be all noise. That's fine ... but it's pretty clear that's exactly what's going on in almost every single internet debate on investing.

jacob wrote:That is to say any given investor is willing to agree that there's more to it than just luck exactly up to the point/level of nuance they personally understand ... beyond that point, it's all luck as far as they and everybody else is concerned. Much like resolving a fractal and declaring any detail beyond one's personal limit of resolution to be all noise. That's fine ... but it's pretty clear that's exactly what's going on.

You're free to think so, but there's so much evidence that most of those who try to beat any given market end up doing worse that I can't even be bothered to find a link. This means that if you pick a random person who thinks he can beat the market then the odds are he is wrong. There are of course exceptions, but it's very hard to know who they are.

@Scrubby - I appreciate you taking it to the next level of regression. Similarly, it must mean that if you pick a random person who thinks he understands how statistics proves that beating the market is impossible then the odds are that he is wrong as well, because the fraction of people who actually understand statistical arguments is much lower than the fraction of people who think they do

The rest of the chain of arguments is proof by induction. Left as an exercise for the reader.

jacob wrote:Similarly, it must mean that if you pick a random person who thinks he understands how statistics proves that beating the market is impossible then the odds are that he is wrong as well, because the fraction of people who actually understand statistical arguments is much lower than the fraction of people who think they do

Sure, if I decided to ignore peer reviewed research and trust a random know-it-all on the internet.

Random. You keep using that word, I do not think it means what you think it means.

Being a non-random guy who has actually worked in finance on matters where a good understanding of randomness is crucial (options, no less), I know that peer-reviewed finance publications lag 15-20 years behind practitioner knowledge because of lack of data (financial data is very expensive and quite beyond what a university grant can afford) and the fact that practitioners who are still making money are not-so-surprisingly unwilling to disclose how they're making money and suffer peer-review until their strategy no longer works. If you want to know, join a financial company. It'll probably blow your mind!

Really, is it hard to believe/accept that times have/could have changed in finance since the 1970s?! Do you think academia is the end-point for all knowledge for all things?

If you base you conclusions on trust instead of critical thinking, maybe trust someone who has worked in finance and is now able to talk about it w/o being restrained by non-disclosures, rather than what makes it into textbooks a few decades later. Forsooth ... the discovery of non-randomness is so old by now that it's even making it into textbooks. The Nobel prize in economics for 2013 should be a giant hint.

BRUTE wrote:so isn't the conclusion that all humans will be disappointed, because nobody had a crystal ball? the only human who will have had success will have gotten lucky. even the humans that "bought the market", they just got lucky with their definition of "the market".

I think you have this right now. You can't use historical data to pick the perfect portfolio. Even determining what is the market is extremely difficult because how much of the market is bonds or commodities or stocks ?

We all have to take a step back and follow some general principles but use some common sense. The people who micropick an asset allocation and think they are going to out perform I think have just put the odds against them.

Some general guidelines though are worth sticking with:-

1. Diversify across asset classes.
2. Diversify within asset classes.
3. Obtain the lowest fees possible.

You also though need to look at the asset classes that you are utilising and view their positive and negative points rather than look historically at the data and think they will always function as they have in the past.

If you follow these principles though you should do okay.

I personally also think that a simple approach works best. It's just less work.

Lastly and most importantly if you have a reasonable sized portfolio (so a 4% WR for instance) and utilise a reasonable asset allocation you should be fine.

@Tyler - I think we are pretty similar in our viewpoints but you have greater confidence in using historical data to determine your portfolio.

I came up with a couple of trading systems myself via using data. I also traded on gut feel. I didn't do well with either approach but gut feel worked a lot better. I'm not stating data mining can't work but in my experience it's not as good as doing research and then making an informed choice. Data is just part of that research.

Scrubby wrote:

jacob wrote:That is to say any given investor is willing to agree that there's more to it than just luck exactly up to the point/level of nuance they personally understand ... beyond that point, it's all luck as far as they and everybody else is concerned. Much like resolving a fractal and declaring any detail beyond one's personal limit of resolution to be all noise. That's fine ... but it's pretty clear that's exactly what's going on.

You're free to think so, but there's so much evidence that most of those who try to beat any given market end up doing worse that I can't even be bothered to find a link. This means that if you pick a random person who thinks he can beat the market then the odds are he is wrong. There are of course exceptions, but it's very hard to know who they are.

@Scrubby - I think that this is the truth.
@Jacob - I don't believe that you are grasping the concept that you are trying to predict the future. You've basically constructed a meaningless argument. It has nothing at all to do with people's understanding or lack of understanding. I think you can put the odds in your favour up to a certain point but I personally don't see that the potential increased returns along with the associated potential decreased returns combined with the additional effort is worth your while. If you think differently that is fine but it's not an understanding thing.

Yeah, uh huh... #dunning-kruger #curseofknowledge #micdrop

steveo73 wrote:The people who micropick an asset allocation and think they are going to out perform I think have just put the odds against them.

out perform what?

@jacob: brute finds it somewhat hilarious that after a PhD in physics, and subsequently switching into the industry side of finance, humans still can't believe jacob knows better

maybe jacob could be somewhat more specific and give some ideas/examples on how humans could positively impact their risk or return. because as brute sees it right now, everything any investor has ever done was inconceivably random.

@brute - Inconceivable, yes! Indeed, ... Obviously you grokked the meme/reference.

So step 1: Acknowledge what might be currently "inconceivable to you" is actually and potentially conceivable with personal effort, e.g. realizing that there's a difference between known-unknowns and unknown-knows. This realization, as far as I've noticed, is absolutely the biggest difference between those who positively impact return/risk ratios compared to the cult of ignorance. Indeed, it's the difference between personal growth and stagnation for all subjects of knowledge. I find that ERE or FIRE itself is not that much different from investing when it comes to one's frame of mind wrt a particular subject. There will always be people who think that whatever subject they don't personally understand is impossible to understand for anybody(!) And because of the internet they will always be able to google some experts (academics or gurus or whatever) to "prove" it too whenever it comes to subjects that don't have an immediate answer to hit people on the head with (e.g. tobacco, global warming, national debt, sugar consumption, meat eating, the benefit of exercise, personal finance choices, ... ). However, this behaviour is a dead-end in terms of personal and intellectual growth. This is the most important thing to realize/accept.

So there's a specific effort for anyone .. get beyond that mental mode. Just because you don't see it(*) doesn't mean it isn't there,



(*)The cigar in the wall, that is!

Because I find these "learning-level" discussions frustrating rather than hilarious, I would also say, that if my induction argument above wasn't obvious, don't bother to waste you time levelling up/going meta, rather just copy the prevailing paradigm and be done with it. This is not for everyone. This is totally fine. I recommend the standard approach to most people including friends and family.

Note: I do recognize that repeating "randomness as gospel" has memetic value because it reinforces ignorance, which reinforces faith, which again supports staying power and the survival of the strategy, much like religion, but on the other hand ... eventually, if a strategy relies on simple reinforcement without support based on fundamentals, it is not self-consistent and it will break eventually (the link is the best investment metaphor I've ever seen). But maybe it'll work long enough and adherent will [unknowingly] adapt to another one. So there's that. It takes a certain time to punish ignorance. I heartily recommend "whatever current paradigm as supported by academic teaching to undergraduates" to anyone who doesn't want to bother.

I think what we have is very much a competition between supporting the faith of the cult of ignorance vs inspiring the minority who are actually able to make a difference for their return/risk numbers. I get vocal when the former discourages the latter.

So step 2: Study and keep at it. I've mentioned where to start before here and there. It's quite evident that the "random crowd" hasn't gotten much beyond a few anecdotes/popular books. Anyone who is really clueless as to where to start, first go figure what your average CFA guy has read. Then read that! It's only 10000 pages. Then PM me.

The end goal should be to sort yourself (you personally) into one of the following four groups:
1) Those who know that they can beat the market.
2) Those who know that they can't beat the market.
3) Those who don't know yet that they can beat the market.
4) Those who don't know yet that they can't beat the market.

I find that group 3 and 4 are citing group 2 (academics and failures) as ultimate proof annoyingly often. This is to the advantage of group 4 and to the detriment of group 3. Yes, if you don't want to bother, you can punt and join group 4 but please don't be an ass and join the chorus of group 2 and 4 preventing group 3 from joining group 1, eh? Fair enough?

jacob wrote:

Tyler9000 wrote: However when you look at all of the returns as a whole, a pattern emerges. All asset allocations are inherently unpredictable, but some are less so than others. By understanding the nuances of uncertainty rather than running from it, one can select a portfolio that clearly has a higher probability of achieving your goals. The differences in average and standard deviation (spread) are inherent qualities of a particular portfolio and are not simple blind luck. Asset allocation matters.

Yes that (my bolding)... but also to repeat this insight when one substitutes "asset allocation" methods with "stock picking" methods: Here's Buffett http://www8.gsb.columbia.edu/rtfiles/cb ... tt1984.pdf ... using pretty much the same argument you made but for stocks instead of asset classes. Like "how all stock picks are inherently unpredictable but some are less so than others". ... "That it's impossible to pick the one best portfolio but that specific classes of picking have differences in average and sds according to the inherent quality of the investment method and thus aren't simply blind luck".

I can see that. I fully admit I understand the mechanics of asset allocation better than stock picking, but I can imagine that once one masters the particular medium similar brush strokes may apply.

Obviously sustainable active management is not easy. Most people would likely be better off financially by picking a passive plan and focusing their energy on maximizing their career earnings and minimizing their expenses. But I acknowledge that just because something is difficult does not make it impossible, and I certainly don't fault anyone for putting their God-given talents to good use.

Ultimately we all have to find the method we're comfortable with and be happy -- both for ourselves and for others.

Hmm, I am both FI and non-FI. As an individual I am FI with more than a 500K AUD stash, however as a couple we are not FI since my wife insists on having a ROI of 3K per month. Together we are close to 700K AUD, but other than money in investments we need around 100K for land and a house to build in Thailand. So another 2-3 years to go. Four years ago I thought I would be FI with 500K, but the goalposts keep moving out of reach. Inflation is to blame as well.

BRUTE wrote:

steveo73 wrote:The people who micropick an asset allocation and think they are going to out perform I think have just put the odds against them.

out perform what?

JamesR wrote:

JL13 wrote:What about the reverse? I'm quitting with a 7.5% withdrawal rate.

Me too probably. I think something like 5-6% SWR is totally valid based picking out an excellent portfolio allocation that does better than 7-8% CAGR on portfoliocharts.com and a greater willingness to deal with down years w/ smaller withdrawals + part-time work or some variation. I agree with MMM's "retire first then get rich" notion, that we're going to still work or produce value on some level anyways even after a super early retirement.

I think I already answered Brute's question but I'll try again. I can't do a comparison of outperformance because I can't quantify what you are comparing outperformance too. I personally don't believe that I will outperform with my asset allocation and I definitely don't believe that my portfolio will give me some special performance based on whatever criteria that you are choosing.

I think that I have a good chance of obtaining a fairly average return. I doubt I will outperform the average portfolio but here is the kicker I also think that I won't underperform the average portfolio and because I won't pay much fees I'll probably come out in front of most people that believe they will overperform.

jacob wrote:Random. You keep using that word, I do not think it means what you think it means.

I ended up on this forum through a series of coincidences. If I really wanted to I could probably dig up thousands of people in forums and blogs with at least the same knowledge and experience in finance as you who would recommend not trying to beat the market. The problem is that you are taking this too personal. It's quite possible that you are one of the few who are able to do it, but that still doesn't make it good advice for others to try.

BRUTE wrote:@jacob: brute finds it somewhat hilarious that after a PhD in physics, and subsequently switching into the industry side of finance, humans still can't believe jacob knows better

maybe jacob could be somewhat more specific and give some ideas/examples on how humans could positively impact their risk or return. because as brute sees it right now, everything any investor has ever done was inconceivably random.

Of course Jacob doesn't necessarily know better than myself. My FIL was Treasurer for Citibank in 2 countries and then managed a billionaire's private hedge fund. Does my FIL know more than Jacob ? Does he know more than me ? Of course he does. That doesn't mean though that Jacob or my FIL can pick a better portfolio than what I can.

With all due respect I trust my own analysis a lot more than other people. I trust that because it's about doing the right thing for myself. I suggest that you do the same thing. Figure out what you are comfortable with.

Do not believe that people are smarter or know better than you when it comes to investing. The truth can actually be the complete opposite. Isaac Newton lost a tonne of money investing.