NEWSFLASH: Bernstein slams early retirement!!!

[unclemick]

Bernstein - all those other guys - should fess up and admit the Norwegian widow was right all along.

Alas - I'm an outlyer at the other end - went up only three fold in the last 12 years. Index funds stuctured to provide 3% current yield. Probably will start drawing some fnds this year.

non cola pension + dividend stocks and and - now at 62 maybe early SS and tapping IRA as needed.

Now about that one stock that will lift me into 'high class ER'

Heh heh heh.l

 

[Mr._johngalt]

Bernstein - all those other guys - should fess up and admit the Norwegian widow was right all along.

Alas - I'm an outlyer at the other end - went up only three fold in the last 12 years. Index funds stuctured to provide 3% current yield. Probably will start drawing some fnds this year.

non cola pension + dividend stocks and and - now at 62 maybe early SS and tapping IRA as needed.

Now about that one stock that will lift me into 'high class ER'

Heh heh heh.l

Unclemick, I 'm not sure "high class" ER would fit your style. Probably not me
either any more. Personally I never doubted the Norweigan widow was
right all along. The DJIA can bounce all over, but those checks in the mailbox
are real money

JG

 

[maddythebeagle]

But for most people, retiring in their mid-30s is going to be risky business and you would have a tremendously hard time finding a financial advisor, or any well known Guru, such as Berstein, who would advocate such an approach.

who needs a guru when you have unclemick? he doenst have to be "lucky", just cash them steadily increasing dividend checks or better yet, get it direct deposited. I plan to have a good chuck coming in that way. Very good plan.

 

[Gone4Good]

It's probably worse than you think it is.  . .   The nine [holidays] we have in total is fewer than any industrialized nation in the world... which to me is simply not acceptable in a country that bills itself as the 'greatest country on earth.'

Earlier in this thread I commented that life is about choices and the consequences of those choices.  This is just as true for countries as it is for individuals.  European citizens work less than their counterparts in the U.S., on average.  The also earn and produce less, on average.  Strict labor laws (like mandatory vacation or enforced 35 hour work weeks) keep the unemployment rate in many European countries at levels that would cause civil unrest in the States. There is no such thing as a free lunch.

People tend to look at one aspect of a complicated situation and pick the one detail that supports the way they want to see the world.  Generally speaking though, if you want to work less you can expect to earn less.  Many people make that trade off all the time, either consciously or unconsciously.  Many participants to these boards took the former route.  I suspect many people who complain they don't earn enough took the later. 

 

[cute fuzzy bunny]

Dang and I thought we were all going to get six weeks to clear brush from now on...

 

[ESRBob]

This really hurts, considering that it's coming from the guy who taught me how to allocate my assets and beat inflation for six decades.

Just back from a week backpacking in the Sierra Nevada -- man I love the ER lifestyle! Don't have the satellite uplink like C-T so have missed out on this long-even-by-our-standards thread.

So Bernstein thinks we should all stay at work? -- guess it explains why he declined to write a blurb for the cover of my book on ER -- Frankly, when it comes to finding someone 'famous' (publisher's request) to write a blurb for a book on ER, you realize that almost anybody famous is working their butt off for a really long time, and has no concept of ER, or particular sympathy for it. So I'd rather by ER than famous, I guess. And I have no interest in keeping up with the Joneses -- have you seen how much they paid for that new BMW?

I do think there is some good news on the SWR front that nobody has hit on yet in this thread-- if you look at almost all the SWR studies, they either take a simple amount in year #1 and adjust it up for inflation every year, regardless of market performance, and/or they use nutty portfolio allocations which have lots of volatility, often involving large percentages of SP500.

If you make a concerted effort to reduce volatility in your portfolio by picking nice low-fee funds from many diffferent equity markets -- small, value, international and oh, sure a bit of those US megacaps which so skew the sp500-- along with bonds (foreign and domestic, high yield and GNMA) and other stuff, possibly outside of funds entirely, (illiquid investments in commercial real estate, commodities, energy, or private businesses), then your SWR calculations will come back very much more favorably than they would from a more volatile portfolio.

Then, if you say 'we all want to stay early retired, so if the market tanks, I'll tighten my belt a little, and maybe even take on a bit of part-time work for some added income", then you can live with an approach that takes a percentage of asset values each year, as opposed to giving yourself steady annual raises despite market performance.

This has the added benefit of giving you a raise if markets go up, something we all like,too.

Actually I modelled all this for my book, with the help of Keith Marbach (zunna.com). We even modelled a rule that lets you take 95% of whatever you gave yourself last year, even if markets tank worse than that, and the cost of this smoothing out was very slight over all the long run (75 year+) data we were using.

Anyway, pull this together and you can be back at your 4%+ withdrawal rate safely over long time periods, with a very high probability of keeping the portfolio intact in real terms, never mind just keeping it from crashing to zero.

So the lesson I take from all this: Keep fishing, keep sailing, keep hanging out in the garden and playing with the kids. Leave the rat race for somebody else.

 

[ESRBob]

Dave,
Exactly -- the thing I try to remember is that SWRs are based around your probability of surviving a relatively rare bad-case or worst-case scenario. Statistically, the future is likely to have even more OK-cases or good cases, and may even have some really-good-cases.

It's one of the reasons I am not in love with the huge TIPS allocation approach -- it effectively lops off all those possible good outcomes in exchange for a guarantee against the relative handful of possible bad cases. I just think there are better ways to manage the downside. Like making small annual adjustments to your withdrawal to keep in step with market performance, or downsizing and belt tightening, or getting a reasonably palatable little source of side-income if need be. All those techniques help you enjoy a higher current income and keep ER from being quite so hard to reach, or quite so spartan once you get there.

 

[wabmester]

Exactly -- the thing I try to remember is that SWRs are based around your probability of surviving a relatively rare bad-case or worst-case scenario.  Statistically, the future is likely to have even more OK-cases or good cases, and may even have some really-good-cases.

Bob, you and I are about the same age.   That means we're both (hopefully) looking at a 50-year withdrawal period.    Statistically, there have only been about 3 non-overlapping 50-year periods in the historical data we all know and love.    I doubt there's one of them that doesn't have a seriously nasty sequence in there.   And, of course, there's only one 50-year period that really matters for us: the one that started a few years ago.

Statistically, the chance that our 50-year sequence is worse that any other in our history is about 1 in 4.    Something for us ER's to chew on.   

 

[haha]

Bob, you and I are about the same age.   That means we're both (hopefully) looking at a 50-year withdrawal period.    Statistically, there have only been about 3 non-overlapping 50-year periods in the historical data we all know and love.    I doubt there's one of them that doesn't have a seriously nasty sequence in there.   And, of course, there's only one 50-year period that really matters for us: the one that started a few years ago.

Statistically, the chance that our 50-year sequence is worse that any other in our history is about 1 in 4.    Something for us ER's to chew on.   

This is the analysis that is most meaningful to me. I am not a security freak by any means, but to me much of the "Don't worry, be happy" talk is no more than wishful thinking.

I also give more credence than many to the meaning of where we are starting from--i.e. valuation.

Ha

 

[sgeeeee]

. . .
Statistically, the chance that our 50-year sequence is worse that any other in our history is about 1 in 4.    . .

Wab,

You've mentioned this before. But what kind of analysis is this 1 in 4 probability based on? I can't imagine any meaningful way to evaluate the probability that the next 40 to 50 years is likely to be worse than any 40 to 50 year period of the past 130 years.

 

[sgeeeee]

. . . I also give more credence than many to the meaning of where we are starting from--i.e. valuation.

Valuation sounds like a reasonable thing to be worried about. I mean, how can value not be important? The trouble is that the way we quantify valuation does not provide any worthwhile meaning.

The valuation an investor cares about is current price-to-future returns. But quantified valuation is always current price-to-past returns. The two are equivalent only if past returns are strongly correlated to future returns. Sometimes they are. Sometimes they are not.

If P/E was really a decent predictor of future returns, there would be a lot more ultra-wealthy investors in the world.

 

[wabmester]

You've mentioned this before.  But what kind of analysis is this 1 in 4 probability based on?  I can't imagine any meaningful way to evaluate the probability that the next 40 to 50 years is likely to be worse than any 40 to 50 year period of the past 130 years. 

If you have 4 horses in a race, what's the probability that any given one of them will be the slowest? Yup, 1 in 4.

200 years of data will mean 4 non-overlapping sets of 50-year epochs. Sure, you can slide a 50-year window around 150 different ways if you want to, but that just means that the next 50 years will be partitially included in 50 of those 150 sets, which gives you 1 in 3 odds for some part of the next 50 years being part of the worst 50-year sequence.

Now, if you want to feel better, you can interpret that to mean that there's only a 1 in 150 chance that the window will fall precisely on the next 50 years, but it won't feel very good if any part of the next 50 years is part of that worst-case sequence, especially given that the last 20 years or so have been the best 20 years ever.

 

[sgeeeee]

If you have 4 horses in a race, what's the probability that any given one of them will be the slowest?   Yup, 1 in 4.

200 years of data will mean 4 non-overlapping sets of 50-year epochs.   Sure, you can slide a 50-year window around 150 different ways if you want to, but that just means that the next 50 years will be partitially included in 50 of those 150 sets, which gives you 1 in 3 odds for some part of the next 50 years being part of the worst 50-year sequence.

Now, if you want to feel better, you can interpret that to mean that there's only a 1 in 150 chance that the window will fall precisely on the next 50 years, but it won't feel very good if any part of the next 50 years is part of that worst-case sequence, especially given that the last 20 years or so have been the best 20 years ever.

Thanks for the answer. I don't find the analysis to be very reasonable, but I know where you are comming from now. A requirement of non-overlapping intervals doesn't really make sense if the goal is to estimate the worst case interval result.

If I were flipping a coin and wanted to estimate the minimum number of times I would flip tails in a row out of 50 flips, I would not need to use independent trials of 50 flips. Two hundred flips would give me 100 series of 50 flips and a good estimate of the worst case probability. In fact, if I only considered the 4 independent trials represented by the 200 flips, I would likely underestimate the worst case probability.

Investment results are a little bit different than a coin toss since this year's results are related to last year's in some complex and unpredictable way. But the time correlation of stock, bond and inflation performance clearly diminishes rapidly over several years. An estimate of 1-in-4 probability would be accurate if you believed that performance and inflation this year were strongly correlated to the the performance and inflation of the past 50 years. That seems a bit extreme.

 

[wabmester]

Coin-flipping and counting tails isn't really what we're doing here.    Amplitude matters.   It's more like, "what are the 50 bumpiest miles between LA and San Diego?"   If you have a really bumpy stretch of 10 miles starting at mile 20, your first 30 50-mile segments are all going to seem really bumpy.    It's nothing like looking at 100 independent trips.

So, what happens if you add another 50 miles of road into Mexico?   There's an excellent chance (about 1 in 4) that the new road will be the worst 50-mile stretch yet, assuming the old road doesn't influence the new.

Edit: unfortunately, we know that our government is too deep in debt to build any more roads, and all of the road workers are retiring, so beware of bumps ahead!

 

[sgeeeee]

Coin-flipping and counting tails isn't really what we're doing here.    Amplitude matters.   It's more like, "what are the 50 bumpiest miles between LA and San Diego. . .

I agree. Coin flipping is the opposite extreme from a 50 year correlation requirement. Each coin toss is completely independent from the previous. But this year's performance and inflation is not completely independent of last year's. On the other hand, it's hard to believe that the results of 2005 are really highly correlated to the performance of 1955.

My suggestion is that reality is somewhere between the probability of success listed by FIRECALC (coin toss equivalent) and the 1-in-4 probability you suggest.

Consider the question you pose: What are the 50 bumpiest miles beween LA and San Diego? The best estimate comes by considering every 50 mile section, not by considering only non-overlapping sections. If the worst 50 miles occurs between miles 25 and 75, then a strict non-overlapping evaluation will underestimate the worst case bumps.

 

[BigMoneyJim]

The valuation an investor cares about is current price-to-future returns. But quantified valuation is always current price-to-past returns.

Yes, but there's a 1-in-2 chance that future returns will be worse than past returns.

 

[bosco]

If you have 4 horses in a race, what's the probability that any given one of them will be the slowest? Yup, 1 in 4.

Aside from the comments about the questionability of the accuracy of the "probabalisitc analysis" in this quote, I can't help but point out that you are comparing an equity market from a very young, almost frontier, republic to that of a nation with a much more complex economy that is in the middle of converting to a service-sector and outsourcing model, rather than a natural resources and manufacturing economy.

To claim that the first 50-year sequence has a similar expected outcome as the last period seems very questionable to me. This is apples to oranges. I think overlapping periods are probably more accurate, for this purpose, and maybe a shorter time-frame, counter-intuitively, might be more accurate for our purposes.

 

[ESRBob]

I know the U.S. history is what we've been using in all these long studies so far, but I am itching to get my hands on the new Dimson Marsh international long term data to see how it holds up.

Going forward, I hope none of us is tying our fortunes solely to the U.S equity markets or the U.S. economy.

When you get yourself into lots of different asset classes, currencies and countries, a lot of the bumps in the road get ironed out right there.

For now, I guess I consider that overlapping time periods are just a necessary evil in order to get more data sets -- we simply can't do meaningful analysis on the history we have any other way.

In the end, the midcourse corrections approach -- taking a percent of assets each year and living on it -- seems to me the best insurance for long-run survivability. It really makes a difference, during hard market conditions, compared to the old approach of giving yourself a stead inflation-matching raise every year off an initial base.

 

[retire@40]

Going forward, I hope none of us is tying our fortunes solely to the U.S equity markets or the U.S. economy.

The US economy is pretty much the world market. If we ever fall into a depression, the rest of the world will also fall into a depression. If so, the only safety may be in gold.

But, as I've said before, chances of another depression are slim.

 

[ESRBob]

The US economy is pretty much the world market. If we ever fall into a depression, the rest of the world will also fall into a depression.

It has no doubt been true in recent decades. I honestly wonder if it will continue to be true. The domestic demand being fanned up in places like China and India is about to be staggering. Wall St. Journal mentioned last week that China is now the third-biggest market for luxury goods in the world. The sheer numbers of people and their need (and ability to pay for) so many fundamental manufactured goods and ancillary services is so vast that at some point, maybe soon, they won't need the U.S. 'engine' of final demand in order to keep themselves in business.

If that is true, then U.S. global influence may wane, or at least level off, and we may look back on these turn-of-the-century times and talk about the good old days.

Cerrainly China want to stand up and be counted as a superpower -- militarily and economically. And with a billion people more than the U.S. (1.3 billion total), hungry, talented and willing to work hard, they may just pull it off.

 

[bosco]

I know the U.S. history is what we've been using in all these long studies so far, but I am itching to get my hands on the new Dimson Marsh international long term data to see how it holds up.

Me too!

I spent quite a while trying to track it down on the internet to no avail.

If you check out the book on Amazon (The Triumph of Optimists, or something like that) it is 1) expensive, and 2) does not actually contain the data series, just charts that are difficult to convert into data.

PS--don't forget Staunton

If anyone has a line on this data, I would love to hear about it. I'd like to drop it into Gummy-stuff's sensible withdrawal calculator and see what effect it has.

 

[ESRBob]

Ibbotson is selling it.

http://tinyurl.com/e422o

I looked a few months back, and I believe the data carried a hefty fee, which I am hoping will drop or some other way to get access will open up. Maybe someone with academic credentials will get a discount?


EDIT: Shortened URL

 

[unclemick]

Goggle - Angus Maddison - The World Economy.

Data is only from 1890-1992. A truly turgid tome - I decided to pass on that one - let Bernstein peruse and draw conclusions for me.

Heh heh heh

P.S. - Bernstein likes Maddison - and has referrd to his work in past Efficient Frontier articles.