Is There An Unlimited Duration SWR?

[Midpack]

I don't recall a thread on this, though a member who's POV I respect once stated that X% [thought it best to omit the number to avoid bias] is the SWR that conceivably goes on forever. Right or wrong, 4% is the rule of thumb for a 30 year duration, which is not necessarily early retirement - it's closer to standard retirement age (retire at 65 with a 95 longevity plan).

There are a lot of retirement and/or financially savvy people here, wonder what the consensus view might be for this question? Call it 40-50 years from present if you can't get your head around unlimited. And if you can provide some basis for your number, all the better...

 

[youbet]

I think the answer is that you cannot use FireCalc to project past market performance and inflation into an infinite future.

However, you could easily plug in a longer terms, say 50 - 60 or more years, and run FireCalc. If the success rate comes out lower than you're comfortable with, go back and ask FireCalc to calculate what withdrawal rate would survive for 50 yrs at a X% success rate.

Another way would be to just look at the graph. My typical run is 100% successful at 30 yrs. If I change that to 50 yrs, I can just look at the graph and scope out where the first failures occurred.

BTW, I admire your optimism on needing to have your money last for "an unlimited duration." I'm 63 and I'd be delighted if I needed for mine to last for another 30 yrs!

 

[M Paquette]

As I recall, the traditional SWR figure for an endowment is roughly 3%. The investments need to be broadly diversified.

Try Googling up endowment portfolios.

 

[FIRE'd@51]

However, you could easily plug in a longer terms, say 50 - 60 or more years, and run FireCalc. If the success rate comes out lower than you're comfortable with, go back and ask FireCalc to calculate what withdrawal rate would survive for 50 yrs at a X% success rate.

If you did this, you would lose those paths that started in the late 60's and failed in the 30-year runs.

 

[W2R]

About five years ago I read somewhere, from an impeccable source that I cannot remember for the life of me, that 3% will last forever.

This is indelibly emblazoned in my memory. The asset allocation that is required for this result is not, however. Probably around 60:40 or I would have noticed.

So, being cautious by nature, and not being entirely certain of the AA that was used in that study, my answer to your question would be that a 2.5% - 3.0% would last forever. If it was my money, at 45:55 equities:fixed, and it absolutely had to last 50 years, I would feel comfortable with 2.5%.

 

[harley]

I would think some of the Monte Carlo simulations could give you a better answer than a straight historical run. Some of the tests would have to include some worse times than we've actually ever had. I'm not sure if you could get it to simulate a long run of stagnant returns, which is what many are predicting for the near to mid term. But just on a gut feeling, I would think 3%, maybe as low as 2.5% would last as long as you would need. There has to be a way to beat inflation by that much on average, even with relatively passive investing. If things get worse than that in the US, I'd be heading for greener pastures. If none are available, retirement will not be an option anyway.

 

[LARS]

Question: Is There An Unlimited Duration SWR?

Answer: ZERO though I suspect that is not the answer you had in mind. But anything else is based on assumptions. And reality is that even with zero SWR the "value" of your holdings could still lose out to inflation if not invested properly.

Bottomline is there are no certainties in this world when it comes to investments, except maybe taxes.

 

[clifp]

I don't recall a thread on this, though a member who's POV I respect once stated that X% [thought it best to omit the number to avoid bias] is the SWR that conceivably goes on forever. Right or wrong, 4% is the rule of thumb for a 30 year duration, which is not necessarily early retirement - it's closer to standard retirement age (retire at 65 with a 95 longevity plan).

.

It seems to me that only truly safe way to have a permanent portfolio is to keeping spending to a level less than the income earned. So for example if you had 50% Total Bond and 50% Total Stock with SEC yields of 2.84 and 1.45% respectfully you have a withdrawal of 2.15%. The Vanguard managed payout growth portfolio (which is suppose provide permanent capital) right now is 2.5%. Obviously this would mean your income fluctuates a lot. I am sure you could apply some type of floor and ceiling rule, no more than 10% drop or 10% increase annual to smooth the fluctuations.

However over the long term, the underlying income (whether paid out or retained as earning) is key determinate of the withdrawal rate.
So the withdrawal rate is equal to the percentage gain in the gross world domestic product.

 

[LARS]

It seems to me that only truly safe way to have a permanent portfolio is to keeping spending to a level less than the income earned. So for example if you had 50% Total Bond and 50% Total Stock with SEC yields of 2.84 and 1.45% respectfully you have a withdrawal of 2.15%. The Vanguard managed payout growth portfolio (which is suppose provide permanent capital) right now is 2.5%. Obviously this would mean your income fluctuates a lot. I am sure you could apply some type of floor and ceiling rule, no more than 10% drop or 10% increase annual to smooth the fluctuations.

However over the long term, the underlying income (whether paid out or retained as earning) is key determinate of the withdrawal rate.
So the withdrawal rate is equal to the percentage gain in the gross world domestic product.

I think the problem with your analysis is that it forgets about the issue of inflation...

 

[ziggy29]

As the time horizon lengthens, the SWR asymptotically approaches a certain "infinite time horizon" level, but it's not clear or carved in stone where that is. In reality that only works as long as the social, economic and political structures of the places where the money is invested is secure and sustainable.

 

[yakers]

Look up endowment portfolios. Yale is the prototype. If large enough then real diversification is called for including real estate, hedge funds, and foreign investments. AFAIK Yale looks to get 4% to 8% growth forever and safely. But they lost money in 2007/08 too. Whatever magic they have may not be available to folks like us. A friend of mine was once one of several investment managers for the Rothchild family. Their family had a couple hundred year history and is going still going strong. They buy real estate, companies, commodities and make unpublished loans. It works for them and for a long, long time but not much useful there for me.

 

[clifp]

I think the problem with your analysis is that it forgets about the issue of inflation...

Partly true, in theory dividend growth should keep up with if not exceed inflation. So with a more aggressive mix say 75/25 the inflation impact would be pretty minimal (of course the withdrawal rate falls below 2% ). However, you are right bond income isn't going to keep up with inflation.

I think your answer of zero is also wrong. There has been historical huge increases in productivity. For instance the labor required to make a ton steel has fallen by a factor 1,000 since 1920, the labor required to make a transistor by a factor of 1 million since the 1970s. The thing that has enabled these productivity improvements is capital. As suppliers of capital we should get the rewards of productivity.

So I feel pretty confident that real returns of permanent portfolio is equal to improvements in productivity in the world. Of course what that number is historically, much less going forward is matter of great debate. My best guess is greater than 2% and much less than 4%

 

[Independent]

I vote for Zero. The infinite future includes the eventual demise of the human race due to whatever stupid things we eventually do.

Nobody knows the right alpha and beta for an infinite future Monte Carlo run.

If we look at 20th century US stocks, we see a 35 year cycle. If your withdrawal rate was low enough to get you back to your original starting value in 35 years, then that should be enough to fund the next 35, etc. Of course for "infinite" you need 100% probability on those 35 year cycles. Without running it, W2R's 3% seems likely.

 

[donheff]

Take any less than 100% portion of the remaining portfolio each year and it will last forever. If you want one never changing percent, pick any fixed percent of the portfolio and take that every year.

 

[traineeinvestor]

Partly true, in theory dividend growth should keep up with if not exceed inflation. So with a more aggressive mix say 75/25 the inflation impact would be pretty minimal (of course the withdrawal rate falls below 2% ). However, you are right bond income isn't going to keep up with inflation.

I agree with this.

FWIW, my expectation is that our portfolio will have to last 50+ years and my admittedly simplistic and assumption driven plan is that having most of the money in equities and real estate and planning to live off less than the net dividend and rental income is likely to be safe. This suggests taking 2.5-3.0% a year. We will have the usual buffer of cash/near cash to cover the inevitable periods of time when there is a fall in the level of income but very little in the way of bonds - even at 2-3% inflation the real value of those bonds will be decimated over that kind of time period.

I suppose a different approach would be all TIPS and live off the non-inflation adjustment interest component.

 

[SecondCor521]

It depends on when you ask.

After a long bull market, 5-9% seems reasonable.

After a severe recession, 1-2.5% seems reasonable.

This is even if you keep the people you ask the same.

Recency bias.

2Cor521

 

[LOL!]

Can you say S. P. I. A.?

 

[ERD50]

Call it 40-50 years from present if you can't get your head around unlimited. And if you can provide some basis for your number, all the better...

I thought I had some notes on a 'forever' portfolio, but couldn't find them. But that's OK, I did dome fresh FIRECALC runs and probably looked at it in a different light anyhow. Some things are quite interesting (IMO):

First important lesson - Forget about long-term. Thirty, forty, fifty years isn't the problem, things average out by then (historically) - the first 2-20 years is what can put the fear into you.

So the approach I took was to first run FIRECALC for 30 and 40 years with 3.0% and 2.5%. I was looking to see the portfolio maintain its value in the worst case over this time (that should be an indication it could last 'forever'). I always start with $1M so the 'spend' becomes the WR with just a decimal shift. Some numbers:

WR	Years	Lowest Remaing Portfolio
3.0%	30	$568K
3.0%	40	$546K
2.5%	30	$953K
2.5%	40	$1M

So I ruled out 3.0% WR, but 2.5% essentially maintains your original buying power for the 30 and 40 year periods, even in the worst case. Now it gets interesting. I ran the 2.5% for 20 years and the result was a bit of surprise:

WR	Years	Lowest Remaing Portfolio
2.5%	20	$564K

so I ran a bunch more:

WR	Years	Lowest Remaing Portfolio
2.5%	25	$701K
2.5%	20	$564K
2.5%	15	$459K
2.5%	10	$494K
2.5%	7	$556K
2.5%	5	$498K
2.5%	4	$551K
2.5%	3	$559K
2.5%	2	$589K
2.5%	1	$704K

So, can you 'tough out' losing half the portfolio in the first 5 years when you want it to last 40-50 years? Not so easy.

Go ahead and run some more with 2%, with different AAs, but I think you'll find it is tough to get a balance of a low stock % to reduce those dips, and also survive for the 30-40 year periods.


Ahh, I did find this note from a thread I started 3 years ago:

http://www.early-retirement.org/forums/f28/firecalc-dips-in-net-worth-anyone-scared-32866-2.html

A few runs and it seemed to show that if you want some assurance that the buying power of your net worth does not drop by more than half, you are looking at ~ a 2% SWR. At that point, time frame doesn't seem to matter much.

It depends on when you ask.

After a long bull market, 5-9% seems reasonable.

After a severe recession, 1-2.5% seems reasonable.

This is even if you keep the people you ask the same.

Recency bias.

2Cor521

While those numbers may be the "seem reasonable" responses you would get from many, they are exactly the opposite from what they should be. After a long bull, your portfolio is 'pumped up' and ready for a fall - you need to take a low WR at that point. And you can take a higher one after a severe recession.

-ERD50

 

[ERD50]

(quoting myself here): So, can you 'tough out' losing half the portfolio in the first 5 years when you want it to last 40-50 years? Not so easy.

Go ahead and run some more with 2%, with different AAs, but I think you'll find it is tough to get a balance of a low stock % to reduce those dips, and also survive for the 30-40 year periods.

OK, so I ran some 2% scenarios with various AAs - not too much different. Just looking at the 5 year point:

WR	Years	%EQ/Fixed	Lowest Remaing Portfolio
2.0%	5	75/25	$520K
2.0%	5	50/50	$544K
2.0%	5	25/75	$563K

But by the time you get down to just 25% stocks, the portfolio gets hit at the other end - down to $753K and $519K at 30 and 40 years.

Of course, all of this reflects the worst case of history. If we don't hit those scenarios with that timing, things could look rosy indeed. But, we never know....

-ERD50

 

[Gone4Good]

After a long bull, your portfolio is 'pumped up' and ready for a fall - you need to take a low WR at that point. And you can take a higher one after a severe recession.

That is similar to my approach, but instead of messing with my WR, I change my asset allocation. As the stock market continues to rally, my WR declines. And as my WR declines, I don't feel like I should just rebalance into some pre-set asset allocation. I really don't need to take the same degree of equity risk at 3% as I do at 4%, and even less at 2%. And as my WR declines, TIPS look like an increasingly attractive investment alternative. So instead of allowing my equity balance to grow, I reallocate those equity "winnings" to long-term TIPS as an insurance policy. If my account grows big enough, or TIPS yields move high enough, I can see being out of equities all together.

When the next bear market rears its head, maybe I'll consider reversing course. But as things stand right now, my equity allocation has declined by 10 percentage points in the last 3 months. If we have another good year in 2011, I'll probably take it down another 5 to 10 points.

 

[clifp]

Of course all of these FIRECalc runs assume that next 50 or 100 years looks like the last 100+ years in the US.

I think for the future a more global view will be necessary.

 

[Nords]

I think for the future a more global view will be necessary.

I think even Dimson & Marsh avoided speculating on the answer to that aspect of the question. Just when you think you're on the right track, you run across 1920s Germany or 1966 America or 1990s-? Japan...

 

[Gone4Good]

I guess I didn't state in my prior post, although I was thinking along these lines, that 2% probably approaches an unlimited** duration. A 100% 30-yr TIPS portfolio gets you 30 years of 2% withdrawals. There is reinvestment risk at year 30, but at that time you still have 100% of your starting portfolio in real terms and are still only needing to draw 2%.

** For my usage, "unlimited duration" excludes nuclear holocaust, alien invasions, and any, and all, Mad Max type scenarios.

 

[donheff]

All of these Firecalc runs with various starting WRs just seem to show that you can have a really, really bad run for a while. Clearly if you lose 50%+ of your portfolio in the first few years you are in a tough spot. It all just reinforces our discussions that you need to be able to shelter your portfolio from large withdrawals in the event of a crash (particularly in the early years). Two ways we have discussed for doing this are reducing spending or segregating an emergency cash stash apart from the portfolio.

 

[BigNick]

Can you say S. P. I. A.?

Yep... if I were afraid of living 40-50 years past my retirement date, I'd be looking at more annuities.

Paying an SPIA premium is in effect pushing the SWR risk on to an insurance company. I was interested to see here that for a 50-year-old single male, $100K buys you $481/mo for life, which is 5.772%. If we allow 2.5% for inflation, that would suggest that an SWR of 3.2% is possible, even if you have to pay commission and make a profit. (Of course, on the flip side, the insurance company can afford a certain number of "SWR failures"; we typically can't).