Juggling Trinity Study Data AOF Style

[Anne]

This thread is giving me a headache.  Here's my thoughts (which are also of questionable worth):

The Trinity study was based on data from I think 1925-1995.  A period of great economic growth overall (which was coupled with exponential population growth), but that growth has also had huge environmental repercussions.  I can't help but think what Planet Earth will be like if that kind of growth continues for the next 100 years.  I don't think it's sustainable.  I actually don't want that kind of growth to continue, unless it's concentrated in sectors such as renewable energy.  If that type of growth is required to maintain a safe 4% withdrawal rate, I don't want it.

I'd rather focus on building skills and interests that I actually like to use, so that I can continue to do things I like to do until I die (and potentially get paid for them as a side effect), than focus on worrying about whether the 4% rule is safe or not.  It's nice to have enough money to make work optional, but only if you don't have to worry about stuff like this.

[Zaphod]

Anne wrote:

This thread is giving me a headache.  Here’s my thoughts (which are also of questionable worth):

The Trinity study was based on data from I think 1925-1995.  A period of great economic growth overall (which was coupled with exponential population growth), but that growth has also had huge environmental repercussions.  I can’t help but think what Planet Earth will be like if that kind of growth continues for the next 100 years.  I don’t think it’s sustainable.  I actually don’t want that kind of growth to continue, unless it’s concentrated in sectors such as renewable energy.  If that type of growth is required to maintain a safe 4% withdrawal rate, I don’t want it.

I’d rather focus on building skills and interests that I actually like to use, so that I can continue to do things I like to do until I die (and potentially get paid for them as a side effect), than focus on worrying about whether the 4% rule is safe or not.  It’s nice to have enough money to make work optional, but only if you don’t have to worry about stuff like this.

Click to expand...

You wont have to worry about that kind of stuff, it will all be fine. Not without difficulties and being slow to fix 'obvious' stuff, but eventually it will. The growth has not been proportional or concentrated in any sector. It always rotates, and the top sector today didnt even exist not too long ago. There can be lots of growth without similarly sized direct environmental externalities, though there are always some kind of externalities.

For example the sector you've chose will only continue to grow, and while it wont go away completely the fossil fuel industry will scale back enormously over time and thats already starting to show signs of future changes. And as you mentioned growth has slowed dramatically, and our use of resources in an efficient manner increases over time so that gives old planet earth a lot of room to breathe and consolidate.

Never extrapolate todays problems into the future with yesterdays solutions. This is how they knew we would have world wide famine in the 60/70s, but instead we have less farmland per person than ever, etc...

[ReFinDoc]

As always, taxes will  be your largest expense. No one  can project taxes out 30 years of retirement.  Every year the retiree will be presented with unique opportunities. Booked tax losses, side income, medical expenses all combine to make these predictions near impossible.  How much do I pay for hedging strategies against high inflation or economic collapse?

The Trinity study is a mathematical exercise at best. Maybe a vague starting point. The ability to take less out of your portfolio during down years is a powerful financial tool. The usual bear market is 8 months. Warren Buffett says bear markets are where the serious money is made.

[Donnie]

I gave the article a cursory skim, but scrambling historical returns is a bad idea.  In my opinion, it is worse than worthless.  It is actually harmful because it leads to false conclusions.  I already had this discussion with AOF on this forum and am somewhat surprised that a professional (I think) statistician would continue to proceed down this road.

If you are unaware of SORR, this article may open your eyes to it.  If you are even dimly aware of it, do not pay attention to this.

[Hatton]

My thoughts on all of this.

I am old enough to give some historical perspective.  I first started thinking about retirement planning and what my "number" would be in the 90s.  Most financial people that were advising me at the time said withdrawing 8-10% was feasible. The Trinity Study was revolutionary because it made people realize they needed a much larger nest egg before retirement. This thinking is still around if you listen to Dave Ramsey.

The Trinity Study is a good starting point.  It allows you to figure out roughly what you need prior to retirement.  Health Insurance and taxes of course factor in.

I find when I talk with financially illiterate doctors the first problem is they have no idea how much they spend.  They think it is their salary so that they will never retire.

Personally I believe I have over-saved which is a good problem to have.  If I was retiring with a leaner portfolio I would Withdraw 4% of portfolio value yearly in good markets and 3% in bad markets.  It is simple and if you use a percentage of portfolio method you will never run out of money.

[Zaphod]

ReFinDoc wrote:

As always, taxes will  be your largest expense. No one  can project taxes out 30 years of retirement.  Every year the retiree will be presented with unique opportunities. Booked tax losses, side income, medical expenses all combine to make these predictions near impossible.  How much do I pay for hedging strategies against high inflation or economic collapse?

Buffett says bear markets are where the serious money is made.

Click to expand...

Sure, for those with capital to deploy, which does not represent those retired (unless you own a house and are willing to mortgage/heloc). You have to have some money on the sidelines in order to deploy it in a bear market, which means you've had to be missing some upside on the way up. Best to plug your nose, cover your eyes and just be consistent.

[Complete_newbie]

Donnie I'm genuinely curious why you think it's a bad idea to think of permutations of historical returns. Why not do this thought experiment ?

[Donnie]

Because it is an improper use of statistics. Here’s the prior thread where it was discussed.

https://www.whitecoatinvestor.com/forums/topic/actuary-on-fire/

[Zaphod]

The danger is that people dont recognize the thought experiment for exactly what it is and give more weight than the methods/accuracy make appropriate. Taking and weighting the results as more than just, "neat".

[Complete_newbie]

Yea that's possibly true zaphod. I'm most interested academically don't really care much otherwise

Donnie I'll check out that thread. On surface I don't see any issue with doing Monte Carlo.

[Dont_know_mind]

I think AOF wrote some good blog posts on why it is dangerous to be out of the market (in cash) for extended periods. I wish I had known that 20 years ago.

I think this post is interesting and maybe the Trinity results are more robust than I had thought.

But it doesn't address the main issues which are: 1. Are there any situations when I should be largely in cash ? 2. What is the true probability distribution of this instrument (SP 500) at this point in time ?

The problem is not just of sample size. The problem is also that we are inducing the future based on human social behaviour which is affected by human intentionality. Unlike modelling cooking bacon and eggs, here the eggs can decide they don't like bacon and the bacon can walk off because it doesn't want to get cooked.

Say for example you were an investor in the 1940's in Nazi Germany or early communist China or USSR. In retrospect it is obvious you should have adjusted the probability distribution of your returns (and fled the country). But most of the time you are tilting at windmills going to cash and killing your return.

The trinity study seems quite US (SP 500) centric. I wonder what they assume as a safe SWR in Europe, China, Russia or other countries.

[Zaphod]

Dont_know_mind wrote:

I think AOF wrote some good blog posts on why it is dangerous to be out of the market (in cash) for extended periods. I wish I had known that 20 years ago.

I think this post is interesting and maybe the Trinity results are more robust than I had thought.

But it doesn’t address the main issues which are: 1. Are there any situations when I should be largely in cash ? 2. What is the true probability distribution of this instrument (SP 500) at this point in time ?

The problem is not just of sample size. The problem is also that we are inducing the future based on human social behaviour which is affected by human intentionality. Unlike modelling cooking bacon and eggs, here the eggs can decide they don’t like bacon and the bacon can walk off because it doesn’t want to get cooked.

Say for example you were an investor in the 1940’s in Nazi Germany or early communist China or USSR. In retrospect it is obvious you should have adjusted the probability distribution of your returns (and fled the country). But most of the time you are tilting at windmills going to cash and killing your return.

The trinity study seems quite US (SP 500) centric. I wonder what they assume as a safe SWR in Europe, China, Russia or other countries.

Click to expand...

All of this is extremely US centric. You go outside the US and you will find many a time period longer than your investing horizon where no expected positive return could be assumed from historical data. US is an anomaly. Never discount that ever. I couldnt find the graphic but I've seen it displayed nicely in the past. Some markets didnt come back at all. Scary stuff. Diversify.

The US is home to a small population yet a very productive and wealthy nation, and the worlds reserve currency. These aspects wont always be true, but they are very powerful.

[Complete_newbie]

Sorry but I didn't see any thing in the thread you posted that convinced me that bootstrapping historical returns is bad I'm amyway. It assumes no underlying model while Monte Carlo draws from a model. It introduces model bias. We can argue till kingdom come about this. I would agree that bootstrapping does assume there is no correlation year to year or stock returns and they are truly random events ...which yea I don't think so but if anyone can point to a great study that shows correlation then I'd like to read and understand

Either way I found this excersize useful as I said before but I understand underlying assumptions I suppose what zaphod says is true that people may consider this gospel which is not the intended purpose (I think)

[G]

AlexxT wrote:

All of these studies are interesting, but the solution to all of the problems they raise is the same.

When the market is down, spend less.

Bonds are yielding less, and stocks will probably yield less in the future, so expect to withdraw less than 4%.

If this study scares you, then plan on withdrawing less than 4%.

3% is safer than 4%.

2% is safer than 3%.

If you can’t make the numbers work on 4%, or 3%, then work for another couple of years and run the numbers again.

If you’re retired and the market tanks and you don’t have enough money, go back to work.

You’re welcome.

 

Click to expand...

This thread is more entertaining with a big pour of bourbon.  Seriously, try it.

I agree with you, AlexxT.

Also, I am a worrier by nature.  Combine that with the baseline persona of somebody who knows/cares about SWR (and/or looks at a personal finance website on a Fri night) and it is a recipe for working longer than needed.

If fate allows, my bigger risk will be a kid that gets a large inheritance....

And, yes, Vagabond, worse case scenario: I might actually pick up a shift or two out of retirement if the alternative is drinking Jack Daniels (the whiskey equivalent of Alpo).

[Donnie]

Complete_newbie wrote:

Sorry but I didn’t see any thing in the thread you posted that convinced me that bootstrapping historical returns is bad I’m amyway. It assumes no underlying model while Monte Carlo draws from a model. It introduces model bias. We can argue till kingdom come about this. I would agree that bootstrapping does assume there is no correlation year to year or stock returns and they are truly random events …which yea I don’t think so but if anyone can point to a great study that shows correlation then I’d like to read and understand

Either way I found this excersize useful as I said before but I understand underlying assumptions I suppose what zaphod says is true that people may consider this gospel which is not the intended purpose (I think)

Click to expand...

I don’t think you understand what Monte Carlo is.  The issue isn't that AOF is using a Monte Carlo method, it’s that AOF is sampling from an improper probability distribution. Basically the only thing that matters in Monte Carlo is to get the probability distribution right, and AOF certainly hasn’t here.

Of the folks posting in this thread, I can practically assure you that I am the only person who has (i) built a Monte Carlo simulation and (ii) had those results published in Nature.  The analysis is garbage other than to illustrate in a purely hypothetical case that the sequence of returns can matter.  If my other posts didn’t convince, you can just take my word for it.