Trinity Study Authors update their results

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
VictoriaF
Posts: 19508
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Post by VictoriaF »

wade wrote:I survived the CFA exam.
Best wishes for passing the exam,

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)
User avatar
KarlJ
Posts: 342
Joined: Mon Nov 03, 2008 8:41 pm

Post by KarlJ »

William Bengen’s studies have been referenced extensively in this thread, so I thought that those interested could check a recent Forbes article that discusses the current Bengen recommendations, which essentially are a “Safemax” withdrawal rate of 4.5% for a portfolio with a 55% allocation in stocks with the remainder in intermediate-term government bonds.

http://www.forbes.com/forbes/2011/0523/ ... ution.html
User avatar
bob90245
Posts: 6511
Joined: Mon Feb 19, 2007 8:51 pm

Post by bob90245 »

KarlJ wrote:William Bengen’s studies have been referenced extensively in this thread, so I thought that those interested could check a recent Forbes article that discusses the current Bengen recommendations, which essentially are a “Safemax” withdrawal rate of 4.5% for a portfolio with a 55% allocation in stocks with the remainder in intermediate-term government bonds.

http://www.forbes.com/forbes/2011/0523/ ... ution.html
Bengen introduces an untested (and non-specific) wrinkle:
One problem with withdrawing from a portfolio heavily into equities is the corrosive effect of a stock market drop early in retirement--a fall from which the retiree might never recover if he continues to make withdrawals based on 4.5% of his beginning balance. To reduce the chances of this, Bengen, at heart a value investor, looks at current price/earnings ratios and doesn't load up on stocks when they're expensive compared to historical norms--think 2000 and 2007.

Or today. "Buying most American stocks now seems to me like buying a house in late 2005 hoping to get a successful ‘flip,'" Bengen says. "You may end up like the Wicked Witch of the East"--meaning flattened in Oz by Dorothy's falling house.
Ignore the market noise. Keep to your rebalancing schedule whether that is semi-annual, annual or trigger bands.
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Wade, I’m glad you’ve returned to rejuvenate this great thread. I hope that CFA program didn’t drench you in lots of diversionary ivory tower university investment theory. What I like best about this long rich thread, from your start of it right through, is its focus on people’s investment purpose of applying their $$ resources for their future $$-value needs and goals. "Goal-based investing," or goal-focused, I think you called it.

You sure are productive. Do you have, or will you have, more resarch that can be reported within the scope of this thread? (Some of us have stretched this thread's scope a bit here and there.)

KarlJ, I wouldn’t bank on that 4.5% “safe” withdrawal rate that Bengen endorses in the Forbes article you linked us to. While this thread is sure long, it’s rich with material that points well below 4.5%. Wade began with the Trinity 4%, and this thread throws plenty of doubt on that figure – for example, Wades data and graphs on other countries, the inadequate-sample and middle-years-overload flaws in historical-sequence sampling, Monte showing some risk with Wade’s data, Wade pointing out that while his analyses include inflation they omitted fees and taxes, Rodc and others pointing out the great uncertainties in our future-investment-performance assumptions . . . To me, this thread points not up to 4.5% but down toward 3%.

I ran a Monte test of Bengen's 4.5% withdrawals, starting with $1 million, using the same Ibbotson historical data source Wade uses and Bengen says he uses, with Bengen’s 55 / 45 allocation. Here it is, not too good:

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/9 ... tr></table>[/img]

(The key to the symbol for each year’s ending balance is in the graph’s upper left.)

This simulation shows you hit 5% risk of no money in year 22. By year 30 the risk of running out of money is 15%. “Safe”??

That’s with history-based numbers. Now consider Rodc’s well illustrated point about the great uncertainty of those underlying assumptions. And the warnings of Mr. Bogle and many others that the years ahead are likely to deliver returns well below those of history . . .

Mr. Bengen’s 4.5% appears to me a plan for excessive risk of final years in that cardboard box in the alley behind the pawn shop.

Then I encountered something he said that blew away any respect for his analysis. He assured us that if we are worried about living beyond those 30 years, under his plan we have 96% confidence of still having our initial value at the end of the 30 years, to cover years beyond. In the example graphed above, that would be only 4% risk you have less than $1 million at the end of 30 years.

Horsefeathers! That makes no sense!! The Monte graph above shows how far off it is – in that analysis, the risk of less that $1 million left at end of year 30 is 43%. Bengen's 4% risk for this just makes me walk away from what he's saying.

Dick Purcell
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Sometimes I don't get an email saying the thread has been updated...

Thanks about the good luck wishes for the CFA exam. Dick, though the CFA curriculum is full of topics you don't like, I think you will be pleasantly surprised that it does include some matters close to your heart as well.

It was nice to see the article about Bill Bengen, since I like his style of analysis better in terms of just focusing on the SAFEMAX. Dick, this Forbes article doesn't make clear, but I think Bengen is also incorporating some small-cap stocks to come up with his 4.5% number. Large-cap, small-cap, and intermediate-term government bonds. About the initial wealth remaining after 30 years in 96% of cases [there's that 96% number showing up again], I should have the capabilities to double check that using the data he has in mind. The Forbes article doesn't mention the asset allocation for this, but I bet I can figure it out by looking in his 2006 book when I'm in my office. I can post the parameters here as well so you can compare with Monte Carlo.

At any rate, I think that Bill Bengen falls into the DRiP Guy school of thought regarding safe withdrawal rates. He says in the article, "My research was based on reconstructing actual investment results from periods such as the Great Depression, World War II and the inflation-riddled 1970s. Is today's environment tougher than those times? Maybe, but that's not certain."

Okay, but just be careful and remain flexible. But I've been thinking as well, for the 2000 era retirees, if we can make it through the next few years without too much more stock market drama, we may be getting far enough away from their retirement date that things will work out okay. The first 10 years of retirement are vital to knowing the final outcomes.

As for what I will work on next, after I get caught up on a bunch of other work things, I'm trying to decide from among 3 general areas:

- some issues related to variable withdrawal rates

- some simulations for goals-based investing

- returning to my half-finished valuations papers

Right now, I'm feeling most excited about goals-based investing, though this changes on a daily or even hourly basis.
User avatar
VictoriaF
Posts: 19508
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Post by VictoriaF »

wade wrote:As for what I will work on next, after I get caught up on a bunch of other work things, I'm trying to decide from among 3 general areas:

- some issues related to variable withdrawal rates

- some simulations for goals-based investing

- returning to my half-finished valuations papers

Right now, I'm feeling most excited about goals-based investing, though this changes on a daily or even hourly basis.
Hi Wade,

Your statement above has triggered a burning question. Rather than asking it here, I started a thread Keeping track of ideas and references. If you have a chance, please take a look at that thread and respond if don't mind.

Thank you,

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Hi,

Victoria, I will write on your new thread later.

I've confirmed that Bengen is right in terms of his statement in the Forbes article.

I don't know exactly what asset allocation he has in mind, but looking at his book and reconciling it with the Forbes article, I use:

40% large-cap, 15% small-cap, 45% intermediate-term government bonds

In this case, the SAFEMAX falls slightly below 4.5%, so there is one failure. But with a 4.5% withdrawal rate, there are only two years in which the remaining balance after 30 years is less than the initial balance (1968 and 1969). Using $1000 as the initial balance, I will include the remaining wealth after 30 years below for retirees in each year.

Note, like Trinity, Bengen assumes that withdrawals are made at the end of the year rather than the start, which is unrealistic. With withdrawals at the start of the year, there are 3 failed retirements and 6 cases when wealth is less than the initial amount after 30 years (49 / 55 = 89%)

For those following at home, here is the information Dick needs for his Monte Carlo simulations. Asset order is:

Code: Select all

Large-cap stocks, small-cap stocks, ITGB, inflation

Arithmetic means =

   11.8425   16.5701    5.4809    3.0955


Standard Deviations =

   20.5065   32.7897    5.7157    4.1988


Correlations =

    1.0000    0.7936   -0.0116   -0.0014
    0.7936    1.0000   -0.1043    0.0523
   -0.0116   -0.1043    1.0000   -0.0010
   -0.0014    0.0523   -0.0010    1.0000
And remaining wealth after 30 years:

Code: Select all

       1926.00       4560.38
       1927.00       4271.36
       1928.00       2648.88
       1929.00       1587.93
       1930.00       2897.70
       1931.00       4513.64
       1932.00       9879.45
       1933.00       9373.45
       1934.00       4993.53
       1935.00       5303.53
       1936.00       3065.44
       1937.00        956.70
       1938.00       5585.68
       1939.00       3503.15
       1940.00       3411.88
       1941.00       4764.57
       1942.00       8005.69
       1943.00       7763.41
       1944.00       4741.44
       1945.00       3079.97
       1946.00       1865.41
       1947.00       4704.49
       1948.00       5480.23
       1949.00       6182.91
       1950.00       5959.54
       1951.00       5610.40
       1952.00       5334.69
       1953.00       6047.39
       1954.00       7798.99
       1955.00       4424.54
       1956.00       3762.10
       1957.00       4765.25
       1958.00       6384.85
       1959.00       3729.56
       1960.00       4042.17
       1961.00       3954.68
       1962.00       2684.44
       1963.00       4459.51
       1964.00       3259.75
       1965.00       1852.20
       1966.00       1048.25
       1967.00       3491.25
       1968.00        433.78
       1969.00             0
       1970.00       4600.66
       1971.00       5204.61
       1972.00       3923.30
       1973.00       2878.29
       1974.00       8983.74
       1975.00      17449.77
       1976.00      12759.66
       1977.00       9699.52
       1978.00      12001.47
       1979.00      10505.66
       1980.00      10912.32
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Wade –

Thanks for your historical-sequences analysis of the Bengen thing in Forbes, and thanks more for providing information so I could test what you analyzed using Monte. I don't think Bengen said 15% smallcaps in the Forbes article, but used it so my test would be the same thing you tested.

But please read on. I think your confirmation of Bengen’s 4.5% “safe" withdrawal advice in Forbes is premature.

My results using Monte are radically different from Bengen’s and yours using historical sequences. With the numbers you provided, my results using Monte are virtually unchanged from the analysis and graph I showed in a message above – starkly rejecting Bengen.

Here’s a graph of my Monte results using your data and Bengen’s recommended 4.5% “safe” withdrawal rate, starting with $1 million:

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/d ... tr></table>[/img]

It shows you hit 5% risk of running out of money by year 23, and by year 30 the run-out-of-money risk is 12%. “Safe”??

And in flaming contradiction to Bengen’s claim you have 96% confidence of still having your initial value at the end of 30 years -- only 4% risk of less -- my Monte shows the risk of less is 42%!!

Here’s a graph illustrating this result – final balance on the vertical scale (higher is better), probability of doing at least that well on the horizontal scale:

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/y ... tr></table>[/img]

I’ve scrolled up to the ending-balance target of $1 million, my initial balance, and the graph shows 58% probability of having at least that much at the end of the 30 years, meaning 42% risk of having less.

To demonstrate how radically different these Monte conclusions are from the historical-sequence simulations used by Bengen and you, I tested how low I’d have to squeeze my annual withdrawal rate for the 96% confidence of having my starting balance or more at the end of the 30 years. Bengen claims I can have it with a withdrawal rate of 4.5%. My Monte shows I have to lower my withdrawal rate to about 1.5%.

Why such radical differences? I think it is again the same flaw in historical-sequence simulation that warped its conclusion regarding optimal allocations in the Trinity update, discussed extensively in your Trinity udate thread -- historical-sequence simulation grossly over-weights the middle years. In the Trinity update, that flaw led us to over-allocations of stocks and under-allocations of bonds. Here, if I am right, this flaw has the more serious effect of seriously understating the risks of a risky retirement withdrawal rate.

In Bengen's and your historical sequences to test his withdrawal rate, the dismal years of the Great-Depression stock market woes are included in very few of the simulations, and the dismal years of the last decade are included in very few -- but every year in the relatively prosperous stretch from ’55 through ’80 is included in 30 of the simulations. I think this is the principal reason for the radical difference in Bengen's and your results from historical-sequence simulation compared to my results from Monte.

As we’ve noted in other threads, the historical-sequences approach also has the weakness of pitifully few simulations – only two that are non-duplicative, and only 55 that are 29/30ths duplicative of each other. Monte shows that with these magnitudes of standard deviations, we need thousands of independent non-duplicative simulations.

I sure would not rely on the withdrawal advice of Bengen! And more broadly, I am now a more loudly screaming rejecter of historical-sequence simulation, for its triple flaws of (a) way too few samples, (b) those few simulations largely duplications of each other, and (c) overweighting of the middle years. I say don’t rely on Firecalc or the tools of Otar!

For actual decisions on safe withdrawal rate, I’d want to make conservative adjustments of my Monte input assumptions for “expected” return rates (those professors sure can find misleading labels!) – adjust downward below those calculated from history, considering Mr. Bogle’s widely shared warning of low-returns outlook and Rodc’s warning about the statistical inadequacies of our historical bases for our assumptions. And I’d consider further adjustment based on p/e or PE10.

Dick Purcell
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Dick,

Thanks for running these numbers. After accounting for the end-of-year vs. beginning-of-year issue, that success number from the historical simulations drops from 96% to 89%.

But, wow, Monte Carlo shows that number at 58% instead. It's a big difference!

This is important. I'm not sure what else to add right away, but it is important stuff.

Just to be clear, I am definitely not recommending 4.5% as safe, but I was just explaining how Bengen got the number that he got, and that at least it is not a mistake in the context of his methodology.

Now we have to figure out which methodology is best.
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Wade –

Here’s another view of the allocations you described for the Bengen plan – showing ranges of variations from plan in number of years and withdrawal rate.

Look first at the red curve, which represents Bengen’s recommended 4.5% withdrawal rate, $45,000 from the portfolio that started at $1 million. The red curve shows how probability of meeting the withdrawals declines as years go by -- out to 10 years beyond the planned 30.

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/d ... tr></table>[/img]

Fred’s wife Sonja is a little younger, 55 at start of the 30 years, so her age is shown under the horizontal years axis.

By year 30, the red curve shows, the probability of having money left for the withdrawals has declined to 88%. And by year 40, when Sonja reaches age 95, probability of money for the withdrawals has declined to 77% -- 23% risk of no money.

The green curves show the declines in probabilities for other withdrawal rates, labeled in terms of dollars in the box in the graph’s top right. Top label corresponds to highest curve, etc.

The top curve represents 3.5% withdrawal rate ($35,000). At that rate, probability of meeting all the withdrawals through the 30 years is a much higher 97%. And as shown by where I have scrolled on the graph, at that 3.5% withdrawal rate the probability of meeting the withdrawals for five more years, out to year 35 when Sonja reaches age 90, is still a high 95%.

Considering serious probabilities of Sonja or Fred or both of them living past year 30, 3.5% withdrawal sure looks better than Bengen's 4.5% to me.

Dick Purcell
User avatar
KarlJ
Posts: 342
Joined: Mon Nov 03, 2008 8:41 pm

Post by KarlJ »

KarlJ, I wouldn’t bank on that 4.5% “safe” withdrawal rate that Bengen endorses in the Forbes article you linked us to. While this thread is sure long, it’s rich with material that points well below 4.5%. Wade began with the Trinity 4%, and this thread throws plenty of doubt on that figure – for example, Wades data and graphs on other countries, the inadequate-sample and middle-years-overload flaws in historical-sequence sampling, Monte showing some risk with Wade’s data, Wade pointing out that while his analyses include inflation they omitted fees and taxes, Rodc and others pointing out the great uncertainties in our future-investment-performance assumptions . . . To me, this thread points not up to 4.5% but down toward 3%.
Thanks to Dick Purcell and Wade for running the tests on the Bengen portfolio SAFEMAX withdrawal rate referred to the linked Forbes article. This thread has incredible quality and depth, especially the Monte Carlo and historical sequence simulations relating to the important subject of SWR, along with the qualifications concerning the strengths and weaknesses of the different simulation approaches.
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Earlier we had an issue in which historical simulations based on beginning of period withdrawals showed that in 89% of cases, the wealth remaining after 30 years exceeded the initial wealth at retirement. Then, Dick found that this number is 58% with Monte Carlo simulations.

I'm relieved to say that there is no inconsistency here.

I usually work with real data, and I usually look only at MWRs, so that when I discuss remaining wealth with Trinity/Bengen style studies, I often forget to convert to real wealth rather than using nominal wealth.

In the Forbes article, Bill Bengen is making the same mistake. He is correct about his nominal wealth figure, but for monetary amounts after 30 years, he should definitely be talking about real wealth remaining. And Dick's analysis was for real wealth, not nominal wealth.

Actually, if you would agree with me that it is better to be looking at real remaining wealth, rather than nominal remaining wealth, then please also realize that Tables 3 and 4 from the updated Trinity study also suffer from this issue, as those are nominal values.

I re-ran the numbers in order to show both nominal wealth and real wealth after 30 years. Initial wealth is 100 now in this case. For the beginning of period withdrawals, which is what Dick is testing and which is more realistic, the percentage of cases in which real wealth exceeds the initial wealth after 30 years is 55%. This compares to Dick's Monte Carlo estimate of 58%. These are close enough together that they tell a consistent story.

Historical simulations do have the bias against bonds, but I feel much better now knowing that they are not really providing a different answer about this issue.

Code: Select all

Withdrawals at Beginning of Period
Inflation-Adjusted 4.5% Withdrawal Rate
40% Large-cap stocks, 15% Small-cap stocks, 45% Intermediate Term Gov't Bonds
Columns:
1) Year
2) Nominal Wealth remaining after 30 years (Initial Wealth=100)
3) Real Wealth remaining after 30 years (Initial Wealth=100)


       1926        414.94        277.03
       1927        387.17        247.56
       1928        233.01        141.60
       1929        125.04         73.95
       1930        249.48        145.65
       1931        399.23        215.85
       1932        897.87        436.30
       1933        839.90        361.72
       1934        415.99        177.12
       1935        442.37        189.89
       1936        226.32         98.17
       1937         30.22         12.84
       1938        462.73        196.66
       1939        252.04         99.45
       1940        251.90         93.22
       1941        379.14        134.30
       1942        695.12        261.37
       1943        680.25        270.30
       1944        403.60        152.11
       1945        255.03         87.48
       1946        128.65         42.16
       1947        404.35        149.38
       1948        487.77        183.98
       1949        555.66        197.43
       1950        526.59        162.14
       1951        489.89        141.97
       1952        468.75        132.01
       1953        525.29        143.69
       1954        683.47        181.26
       1955        356.31         90.45
       1956        274.15         67.31
       1957        361.81         90.34
       1958        522.14        128.63
       1959        260.54         62.56
       1960        273.04         63.59
       1961        270.22         60.19
       1962        120.11         26.13
       1963        280.36         60.01
       1964        153.05         32.41
       1965         26.66          5.56
       1966             0             0
       1967        131.66         27.33
       1968             0             0
       1969             0             0
       1970        186.13         41.69
       1971        264.92         60.54
       1972        153.53         35.71
       1973         79.94         18.78
       1974        650.87        163.26
       1975       1487.49        405.45
       1976       1061.08        299.28
       1977        770.41        222.11
       1978       1003.15        296.69
       1979        905.37        291.66
       1980        958.70        339.79

Portfolio Success Rate: 95%
Percent of Cases, Nominal Wealth After 30 Years Exceeds Initial Wealth: 89%
Percent of Cases, Real Wealth After 30 Years Exceeds Initial Wealth: 55%
Median Remaining Nominal Wealth: 379%
Median Remaining Real Wealth: 134%


Withdrawals at End of Period
Inflation-Adjusted 4.5% Withdrawal Rate
Columns:
40% Large-cap stocks, 15% Small-cap stocks, 45% Intermediate Term Gov't Bonds
1) Year
2) Nominal Wealth remaining after 30 years (Initial Wealth=100)
3) Real Wealth remaining after 30 years (Initial Wealth=100)

       1926        456.04        304.47
       1927        427.14        273.11
       1928        264.89        160.98
       1929        158.79         93.91
       1930        289.77        169.18
       1931        451.36        244.03
       1932        987.94        480.07
       1933        937.35        403.68
       1934        499.35        212.62
       1935        530.35        227.66
       1936        306.54        132.96
       1937         95.67         40.64
       1938        558.57        237.39
       1939        350.31        138.22
       1940        341.19        126.27
       1941        476.46        168.77
       1942        800.57        301.01
       1943        776.34        308.49
       1944        474.14        178.70
       1945        308.00        105.65
       1946        186.54         61.14
       1947        470.45        173.80
       1948        548.02        206.71
       1949        618.29        219.68
       1950        595.95        183.50
       1951        561.04        162.59
       1952        533.47        150.24
       1953        604.74        165.43
       1954        779.90        206.84
       1955        442.45        112.32
       1956        376.21         92.36
       1957        476.52        118.98
       1958        638.48        157.29
       1959        372.96         89.55
       1960        404.22         94.15
       1961        395.47         88.10
       1962        268.44         58.41
       1963        445.95         95.45
       1964        325.98         69.03
       1965        185.22         38.66
       1966        104.83         21.75
       1967        349.12         72.46
       1968         43.38          9.12
       1969             0             0
       1970        460.07        103.04
       1971        520.46        118.94
       1972        392.33         91.25
       1973        287.83         67.62
       1974        898.37        225.34
       1975       1744.98        475.63
       1976       1275.97        359.89
       1977        969.95        279.64
       1978       1200.15        354.95
       1979       1050.57        338.43
       1980       1091.23        386.76

Portfolio Success Rate: 98%
Percent of Cases, Nominal Wealth After 30 Years Exceeds Initial Wealth: 95%
Percent of Cases, Real Wealth After 30 Years Exceeds Initial Wealth: 69%
Bongleur
Posts: 2276
Joined: Fri Dec 03, 2010 10:36 am

Post by Bongleur »

Remind me why you have to have at least as much as your starting wealth to succeed?

I think its more useful to use a 40 to 50 year withdrawal period and make sure you end up not less than zero.

Another real-world structure for the data is to determine what minimum real interest rate succeeds and then program the backtest & Monte Carlo to tell you how often that minimum rate does not occur.
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

No, it's not that you need your initial wealth to succeed. It's just that I was troubled by what seemed to be a different conclusion between historical simulations and Monte Carlo.
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Wade, that’s great! That’s more valuable work you’ve done, persevering to trace down the explanation for the reported difference in answers. And I know you’re glad, as I am, that it’s just omission of the inflation adjustment rather than a gross inadequacy in (ahem) one of the simulation methods.

The methods do however still show some difference in the original purpose of the analysis – probability of making the 30 years with Bengen’s 4.5% withdrawal rate, which he proclaims “safe” based on historical-sequence simulation. What you’ve just reported shows that historical-sequence simulation does indeed show that withdrawal rate pretty “safe,” only 5% chance of running out of money. But my Monte shows it less safe, 5% chance you run out of money back in year 23 and 12% you’re penniless by year 30.

Bongleur – The only reason I and then Wade addressed that idea of having the original value at the end of the thirty years is that in the Forbes article Bengen made an erroneous claim that his 4.5% withdrawal rate gave us 96% probability of doing so. That was his answer to the question of outliving the 30 years in his analysis.

I agree with you that a better test is to test probability of success for a time horizon longer than 30 years. Indeed, the last graph I posted just a few messages up, with one red curve and other green ones, does exactly that and more. That graph shows success probabilities for various retirement lifetimes of up to 40 years -- with each of several withdrawal rates. In other words, that graph is a whole field of what-if combinations of withdrawal rate and number of retirement years, all compared in success probability visually!

The red curve represents Bengen’s 4.5% withdrawal rate, and the graph shows that by year 40 the probability of success is down to about 76% -- 24% risk of running out of money and moving into the cardboard box. The top green curve represents the more responsible 3.5% withdrawal rate, which even after 40 years gives you about 94% chance of success, only 6% risk of running out of money.

Note however that in that graph as shown in the message above, it’s all based on return-rate probabilities calculated from history. Considering the warning of likely modest return rates ahead from Mr. Bogle and others, and Rodc’s eternal-truth warning that our return-rate-probability data calculated from history have big statistical margins of error, I’d advise reproducing this graph with the asset classes’ “expected” return rates reduced a notch below the numbers calculated from history.

From everything I’ve seen in this long thread, I think the safe withdrawal rate is no higher than 3% -- and that rate should be adjusted up or down based on retirement-day p/e or PE10.

Dick Purcell
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

Bongleur
Posts: 2276
Joined: Fri Dec 03, 2010 10:36 am

Post by Bongleur »

Above was from Feb: 4% safe withdrawal amount

More recent:

Valuation-based market timing with PE10 can improve returns?

http://www.bogleheads.org/forum/viewtop ... &start=200
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
Bongleur
Posts: 2276
Joined: Fri Dec 03, 2010 10:36 am

Post by Bongleur »

It just occurred to me that you could plan to die in debt. I wonder if you could get enough credit cards to fund 3 years or so of living expenses...
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Our discussion on this page has been featured in Michael Kitces' new blog entry:

http://www.kitces.com/blog/archives/169 ... rizon.html
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

I just sent an Email to Mr. Kitces. I think he would be a good "expert" addition here. Just IMHO.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 4:54 pm

Post by DRiP Guy »

wade wrote:Our discussion on this page has been featured in Michael Kitces' new blog entry:

http://www.kitces.com/blog/archives/169 ... rizon.html
Nice article by Kitces.

And it is interesting the various ACTUAL thresholds and provisos and options and 'switch-points' that even the most staid boglehead, or most wild market timer actually set for their individual portfolios. I think there are likely as many of those flavors as their are individual investors!

For me, I will share --

* Yes I am a Boglehead in my investing habits, and have been and intend to remain so. Buy and Hold, conservative AA, fit (and slowly sliding) for my risk tolerance, and fixed-date annual re-balancing for me, all part of my written plan.

* And yes, I very much am an adherent of the "4% seems to be about right" school of allowable SWR, supported by many studies and restudies, including Trinity - both as originally released, and as 'reloaded' and recast.

* However, I may not have been as vocal about a personal 'flexi-goal' that I have. This goal began when I initially met my personal goal of 1MM NW. So, wanting to maintain above that amount could very much impact my DISCRETIONARY spending such that my actual WR may often be less than the allowed 4% if and when my personal net worth balance (investable and real estate) ever should drop to close to that psychologically important but otherwise meaningless one million dollar figure (as measured using mark-to-market for the real estate, for as long as RE is still declining.)

This 'rule' is not part of my written personal investing plan, because it is not a required element for success. It is a personal discipline over and above.

It is, however, something that can and already has influenced my decisions on whether to get another new car, whether to move upstate to a new bigger house with acreage, and whether to spend more or less on my various hobbies. So far, I am fully enjoying an unfettered ER, and my NW remains above the mark, and I think the trajectory is fine, as long as we don't tumble dramatically further, but just continue to sag a bit and eventually (2015-ish?) finally start to recover.

I bet I am not alone in my approach.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

IMHO, the KITCES work is the missing link in SWR studies. But for whatever reason, it has not gotten the attention that it deserves. It solves this very troublesome question: Why can a pre2008 who had $1M in early 2008 take $40,000 per year, yet the 2009 retiree with the same current (as the pre2008 retiree currently has) amount only take (say, for instance) $30,000?

I hope he joins BOGLEHEADS so we can discuss his findings directly with the finder!
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Leesbro63 wrote:IMHO, the KITCES work is the missing link in SWR studies. But for whatever reason, it has not gotten the attention that it deserves. It solves this very troublesome question: Why can a pre2008 who had $1M in early 2008 take $40,000 per year, yet the 2009 retiree with the same current (as the pre2008 retiree currently has) amount only take (say, for instance) $30,000?

I hope he joins BOGLEHEADS so we can discuss his findings directly with the finder!
Leesbro63: I believe you are referring to his May 2008 newsletter.

In fact, Michael Kitces and I are operating on the same wavelength for a number of matters, and I cite his article extensively in my article about safe savings rates from the May 2011 Journal of Financial Planning, as well as in my article, "Can We Predict the Sustainable Withdrawal Rate for New Retirees?" that will be in the upcoming August 2011 Journal of Financial Planning. In the safe savings rate article, in particular, I provide what I think is a reasonable solution for the "safe withdrawal rate paradox" you are describing.

DRiP Guy, it's nice to see you back and active again, and I know you do not believe there is any paradox. I'm becoming more mellow in my views about safe withdrawal rates as well, as obviously people will do what it takes to not run out of money.

Michael Kitces is doing all kinds of interesting things, and I do hope he will join us here!
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

Wade, I used to, but no longer get the "Journal". I hope you will post a link to your article when it's published. I'd also be interested in any other stuff you may have published along these lines. I've been following SWR since Bengen's original 1994 article. (Although my opinion of him has been tainted by his failure to stay the course, required for SWR to work, in 2008.)
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Leesbro63 wrote:Wade, I used to, but no longer get the "Journal". I hope you will post a link to your article when it's published. I'd also be interested in any other stuff you may have published along these lines. I've been following SWR since Bengen's original 1994 article. (Although my opinion of him has been tainted by his failure to stay the course, required for SWR to work, in 2008.)
Hi Leesbro63, for the article coming in August, you can see a draft version now at:
http://ideas.repec.org/p/pra/mprapa/30877.html

For the safe savings rate paper, you can see it here:
http://www.fpanet.org/journal/CurrentIs ... ingsRates/

I think Figure 2 provides an explanation about the paradox, though I didn't discuss it there. I do specifically mention the paradox in the conclusion. Basically, I think the paradox is resolved just by integrating the pre-retirement and post-retirement periods together.

Thank you, Wade
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

I just spent 10 minutes reading your publications. Looks like good stuff but this is gonna require rereading and digesting. Thanks for sharing it. My "knee jerk" observation is that, in theory, the min safe savings rate is a good number, but most people don't focus on it until they are 20-25 years into the work force. Sort of like trying to get teenagers to worry about cholesterol.


By the way, is "Rob Bennett" also known as..."he who cannot be named"?
grayfox
Posts: 5569
Joined: Sat Sep 15, 2007 4:30 am

Post by grayfox »

Mr. and Mrs. Smith are just turning 65 and retiring. During their 40 year working lives they were thrifty and saved diligently and have accumulated a 1 million dollar retirement nest egg. Now they are trying to see exactly how to go about making withdrawals. What should the portfolio hold? How to take withdrawals? How much can they spend each year?

BTW, their living expenses are about $65,000 per year. No pension. Social Security will come to about $20,000 per year, so they need $45,000 from their nestegg.

So they go shopping around for a financial planner.

First, they go one guy, Zvi, who recommends that they invest in 80% in TIPS bonds and 20% in stocks. He tells them that they can withdraw 2.35 % for the next 30 years. So their $1M gives them $23,493 per year in withdrawals. Ouch!

Next, they go to Gray, who recommends that they invest 50% in stocks and 50% in bonds and tells them that they can withdraw 3.13% which gives them $31,284 per year to live on. Still not enough.

Then they go to Michael and he tells them to invest 60% in stocks and 40% in bonds, and that they can withdraw 4.5% per year, which gives them $45,000 per year to live on. And on top of that, there's a 96% chance that they will have 100% of their principle remaining after 30 year. Hallelujah! Finally someone is making some sense.

So who will get the Smith's business?

There is something called confirmation bias, where we only pay attention to results that confirm what we already believe or want to hear.

If you have to tell people that, after saving diligently for 40 years, they can only spend $23,000 per year, you won't have any clients. So you are going to tune out anything that says withdrawal rates have to be lower, and latch on to the anything that says, "No worries."
Last edited by grayfox on Tue Jun 21, 2011 8:33 am, edited 1 time in total.
Bongleur
Posts: 2276
Joined: Fri Dec 03, 2010 10:36 am

Post by Bongleur »

Wade, re the June 12 tabular info you posted...

Suppose you add a 4th column, showing how many more years beyond 30 are supported by the same 4.5% _real_ draw rate, based on each cohort's remaining portfolio amount? It ranges from none to maybe 120 _more_ years, in the case where the ending _real_ value of the portfolio is more than 4x the starting portfolio ??? Something is fishy here... am I reading it correctly?

I wish 30 years was not the default standard timespan for these studies; I think lifespans are such that all work should use 40 years as the amount of time you need to plan for the money to last.

DEF Cohort: each retirement starting year is a cohort.
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 4:54 pm

Post by DRiP Guy »

grayfox wrote:Mr. and Mrs. Smith are just turning 65 and retiring. During their 40 year working lives they were thrifty and saved diligently and have accumulated a 1 million dollar retirement nest egg. Now they are trying to see exactly how to go about making withdrawals. What should the portfolio hold? How to take withdrawals? How much can they spend each year?

BTW, their living expenses are about $65,000 per year. No pension. Social Security will come to about $20,000 per year, so they need $45,000 from their nestegg.

So they go shopping around for a financial planner.

First, they go one guy, Zvi, who recommends that they invest in 80% in TIPS bonds and 20% in stocks. He tells them that they can withdraw 2.35 % for the next 30 years. So their $1M gives them $23,493 per year in withdrawals. Ouch!

Next, they go to Gray, who recommends that they invest 50% in stocks and 50% in bonds and tells them that they can withdraw 3.13% which gives them $31,284 per year to live on. Still not enough.

Then they go to Michael and he tells them to invest 60% in stocks and 40% in bonds, and that they can withdraw 4.5% per year, which gives them $45,000 per year to live on. And on top of that, there's a 96% chance that they will have 100% of their principle remaining after 30 year. Hallelujah! Finally someone is making some sense.

So who will get the Smith's business?

There is something called confirmation bias, where we only pay attention to results that confirm what we already believe or want to hear.

If you have to tell people that, after saving diligently for 40 years, they can only spend $23,000 per year, you won't have any clients. So you are going to tune out anything that says withdrawal rates have to be lower, and latch on to the anything that says, "No worries."
We almost agree.

Rather than being an example of 'confirmation bias' I'd just call that 'smart shopping'!

:lol:

Now, if we have reason to believe that the search for higher SWR (what *else* would anyone do?) leads to specious findings, and/or unstated/undisclosed additional risk, or delusional/incomplete analysis, then you may have a complaint to make.

But outside of that, the beautiful thing about a free market, and a bunch of humans trying to make a better mouse trap is to be able to enjoy the fruits of those labors.

After all, aren't Wade and all the others with their P/E10 and other timing schemes attempting to 'move the mark' higher than traditional methods?

I say now (and always have) more power to them, with just one proviso, one I certainly hope you would agree with -- be honest and direct by detailing your approach's specifics, so that it can be properly tested and evaluated in all critical dimensions (risk, return, costs, difficulty for an individual investor to implement, etc).

Unlike the one recently re-named lone crackpot who I *wish* would have remained nameless (sigh-lol), I do think just about everyone else playing the game of "find the SWR" publicly is playing within those parameters.

And IMHO, we all gain from that, whether we are actually swayed in the end to change our own approach or not.

And I am glad you chose a couple whose nest egg versus desired retirement income was a stretch -- it has been said time and again, and the two corollary statements still bear repeating:

* If you find it takes additional risk over what you anticipated, in order to fund your goals, then you should consider lengthening your accumulation period.

* Take ONLY as much risk as needed to fund your goals. For most Bogleheads, they are pretty conservative, and I'll bet many of them are in all or nearly all fixed instruments merely because that vehicle adequately meets their needs, without excessive market risk exposure.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

The problem might be that they also go see an advisor named Cramer. He tells them with good stock selection, they can withdraw 8%. but since theyre BOGLEHEADS, they realize that Michael/Wade is the reasonable compromise between Cramer and Zvi. They then live comfortably ever after, with occasional pangs of market anxiety,and their kids end up with a nice inheritance too.
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Bongleur –

For “safety” of that Bengen 4.5% at various numbers of retirement years, out to 40, scroll up to my post with the graph showing one red curve and four in green. It shows survival probability on the vertical axis v. number of years on the horizontal, and goes out to 40 years at the graph’s right edge. The red curve represents Bengen’s “safe” 4.5%.

Reading off numbers with that graph’s scrolling, I can report that it shows risk of pennilessness rising to 5% as early as year 23, to 12% out-of-money risk by year 30, to 24% chance you have zero money by year 40.

“Safe”??

That graph is based on the same historical data and allocations Wade reports were used by Bengen. But -- it was produced using Monte Carlo simulation rather than Bengen’s method of historical-sequence simulation. In light of the terrible deficiencies of historical-sequence simulation revealed in the rich discussion of prior pages in this thread, not to mention the questionable assumptions (see Rodc), I’m appalled that people keep promoting that very widely promoted 4% figure determined from historical-sequence simulation as “safe.” Not to mention Bengen’s wildly optimistic 4.5%.

Dick Purcell
Bongleur
Posts: 2276
Joined: Fri Dec 03, 2010 10:36 am

Post by Bongleur »

OK now I know who Kitces is:

Kitces (2008)
http://www.kitces.com/assets/pdfs/Kitce ... y_2008.pdf
demonstrated that "the safe withdrawal rate for a 30-year retirement period has shown a 0.91 correlation to the annualized real return of the portfolio over the first 15 years of the time period".

source: http://advisorperspectives.com/dshort/g ... eturns.php

***

Here's another real world perspective:

If you have the usual theoretical $1M you can take $25,000 for 40 years. That is a real $25k if you only but keep up with inflation. Suddenly those 0% yield TIPS don't look so bad.

And TIPS at the current 1.85% ???

A table showing the annual "40 year real withdrawal" for real portfolio returns from zero to 2% would be very useful. What's the historical average -- 1.5% IIRC ???

Don't know how to account for possible immediate taxation of "phantom income" when unexpected inflation exists. Might require a spreadsheet to punch in variable values, or another table showing the reduction for taxes at various percentages of unexpected inflation. [ yeah yeah you _should_ hold TIPS in tax-advantaged blah blah blah ]

Hmmm, If 10% of your money is tax-advantaged and in TIPS at 2%, then what % of your portfolio is keeping up with inflation ??? Brain is dead, time to go to sleep...
Seeking Iso-Elasticity. | Tax Loss Harvesting is an Asset Class. | A well-planned presentation creates a sense of urgency. If the prospect fails to act now, he will risk a loss of some sort.
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

I’m alarmed that almost everywhere, I keep seeing that 4% withdrawal rate labeled as “safe.” And now even Bengen’s 4.5% in Forbes appears to be getting respect. At risk of some repetition from prior pages of this thread, I want to right here take a sledge hammer to the notion that for retirement-lifetime withdrawals, 4.5% or 4% is “safe.” They are not so safe!

>> Horribly inadequate evidence -- The notion that 4% (or even 4.5%) is “safe” comes from historical-sequence simulation, which has fatal inadequacies. Applied to data from 1926 forward, for 30-year periods, it has woefully inadequate numbers of simulations – a mere 50 or so, depending on final year. But with standard deviations of the magnitudes of stocks, assessment of result probabilities for 30 years calls for thousands of simulations. Even those puny 50-or-so historical-sequence simulations are not separate, but almost entirely duplications of each other. There aren’t even enough years for two of those 30-year historical-sequence simulations that are separate and non-duplicative. And as we discussed extensively in prior pages of this thread, historical-sequence simulation is essentially just a big overload of duplications of the middle years – for example including years 1955-80 in 30 of the 50 simulations, but year 1929 in only four and 2008 in only three. On the most basic statistical criteria, this historical-sequence “evidence” just plain flunks in flying colors!

>> Monte Carlo reveals more risk, less “safe”-ty -- When the same history of the same portfolio is converted into return-rate means and standard deviations and subjected to a more robust 10,000 Monte Carlo simulations, it consistently shows more risk of running out of money, less “safe”-ness. That’s true with Trinity, and true with the Bengen thing in Forbes, as illustrated in prior pages of this thread.

>> Omitting probabilities of longer life -- Those analyses, Trinity, Bengen, and the Monte Carlo mentioned in the paragraph just above, all assume the investor doesn’t outlive the 30 years. Obviously the significant probabilities of a longer life make the plan less “safe.”

>> Omitting fees -- Those analyses assume no investment costs or fees. Obviously with 30 years of compounded effect, costs and fees will reduce the “safe”-ness

>> Omiting taxes – Those analyses all assume no taxes. Whatever taxes the investor must endure will certainly further reduce the "safe”-ness. The kind of Monte Carlo test cited above – assuming you disappear at the end of the 30 years – shows that if fees and/or taxes shave 1% from your return rate, risk of running out of money reaches just about 20%. Did I mention that’s if you don’t outlive those 30 years?

>> Statistically ill-grounded historical probabilities – In their different ways, all of these analyses assume the evidence we garner from that history is the “true” measures of historical probabilities. But as intrepid Bogleheads contributor Rodc is wont to remind us, citing return-rate probabilities used for Monte Carlo to illustrate, for those assumptions we have such an inadequate number of years that our calculated measures may be off the “true” measures of that history by 2% or more. For example, for nominal return-rate mean for large-cap stocks, we might calculate 11% while the “true” is 9%.

>> Assumes future probabilities will be as good as those of that history -- All of this is also assuming that whatever the “true” performance probabilities were for that history, that the same will hold true through the 30 years ahead. Mr. Bogle warns maybe not.

In Michael Kitces’ blog page linked-to in other postings above, he stresses the very valid point that with 4% or even 4.5% withdrawal rate, chances appear very good for results far above what’s needed. But the spread of 30-year result possibilities is very wide. Based on the points above, I’d be wary of labeling as “safe” any withdrawal rate above 3%. Even lower if PE10 is high.

I'd want to test it with Monte rather than historical-sequence simulation, including and testing assumptions for fees and taxes and longer lifetimes, and testing with return-rate means and standard deviations and correlations less favorable than those we calculate from the asset classes' inadequate histories.

Dick Purcell
Last edited by Dick Purcell on Tue Jun 21, 2011 11:53 am, edited 2 times in total.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

People will drop out at SWRs too low. If you tell someone they can only take $30,000 on his $1M and must pay taxes on ALL dividends and interest on the full $1M OUT OF THE $30,000, they'll just give up and quit. Even at 4% it's a bitter pill for all but those of us in the know. That doesn't mean that 4% is safe, but perhaps it is "safe enough" versus the "I'll never have enough so fark it" alternative.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

What is the current PE10?
Yipee-Ki-O
Posts: 167
Joined: Sun May 24, 2009 12:42 pm

Post by Yipee-Ki-O »

Leesbro63 wrote:I just spent 10 minutes reading your publications. Looks like good stuff but this is gonna require rereading and digesting. Thanks for sharing it. My "knee jerk" observation is that, in theory, the min safe savings rate is a good number, but most people don't focus on it until they are 20-25 years into the work force. Sort of like trying to get teenagers to worry about cholesterol.


By the way, is "Rob Bennett" also known as..."he who cannot be named"?
Yes, that would be the one and only Robert Michael "Hocus" Bennett.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Post by Rodc »

Leesbro63 wrote:People will drop out at SWRs too low. If you tell someone they can only take $30,000 on his $1M and must pay taxes on ALL dividends and interest on the full $1M OUT OF THE $30,000, they'll just give up and quit. Even at 4% it's a bitter pill for all but those of us in the know. That doesn't mean that 4% is safe, but perhaps it is "safe enough" versus the "I'll never have enough so fark it" alternative.
The solution is not to pretend that SWR rates are higher. The solution is to abandon SWR altogether as a plan (though ok for a penciled in rough planning number).

One needs a flexible plan if they are going to use an uncertain and time varying income producing portfolio. IMHO.

If you want safety build a solid plan for a baseline steady stream of income to fund what is required. Then you can fund your fun money from an uncertain and time varying income producing (hopefully growing) portfolio.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 4:54 pm

Post by DRiP Guy »

Leesbro63 wrote:People will drop out at SWRs too low. If you tell someone they can only take $30,000 on his $1M and must pay taxes on ALL dividends and interest on the full $1M OUT OF THE $30,000, they'll just give up and quit. Even at 4% it's a bitter pill for all but those of us in the know. That doesn't mean that 4% is safe, but perhaps it is "safe enough" versus the "I'll never have enough so fark it" alternative.
What does 'drop out' & 'quit' mean in this context?

(EDIT TO ADD: Ah, never mind, I think YKO nailed it in a later post! ) :lol:

To answer Dick's point about the historical data being woefully inadequate, well, that's as may be, but it is what we have, that is *real* and experiential. Nothing wrong with Monte Carlo simulations, as long as we realize what they are intended to do and all they can do -- express a range of probable outcomes based on input data.

At the end of the day, taken to the furthest stretch, well, nothing is really safe is it? I mean, we all watched 'Fail Safe" and "Dr Strangelove" when we were younger. The risk of two superpowers duking it out to the bitter end of nuclear Armageddon maybe measurably less today (ah, but we have few historical sequences to compare! Surely, we must break out the Monte Carlo!), but as Robert Frost reminds us in "Fire and Ice" there is more than one way to kill a cat (or a whole planet of cats w/ their associated keepers). So danger abounds, uncertainty exists, and yet still, try we must, and plan we should. I am glad ultra pessimists are out there and I'm glad people like Kitces and others are out there claiming they can rejigger for 4,and 5,and +6% SWRs with their own brand of proposed 'secret sauce'.

I am repeating myself repeating myself, but I figure 4% allus worked, and it's a good enough Kentucky Windage number for my own planning, unless and until the world is really turned on it's ear with some groundbreaking changes. I haven't seen 'em yet, but I do stay tuned.

Really.
Last edited by DRiP Guy on Tue Jun 21, 2011 12:59 pm, edited 1 time in total.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 4:54 pm

Post by DRiP Guy »

Leesbro63 wrote:What is the current PE10?
http://www.multpl.com/
User avatar
1210sda
Posts: 1723
Joined: Wed Feb 28, 2007 8:31 am

Post by 1210sda »

Wade's research paints a tough picture . For 2010 retirees, he cites a MWR of 1.8%. That's 55% below the popular 4%.

Wade states his hope that his predictions for recent retirees are too low. Me too.

1210

P.S. back to the drawing board
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

Yipee-Ki-O wrote: Yes, that would be the one and only Robert Michael "Hocus" Bennett.
Dare I ask whatever happened to him?
User avatar
Dick Purcell
Posts: 520
Joined: Tue Oct 26, 2010 1:58 am

Post by Dick Purcell »

Lee, I agree!

I didn’t mean to say “Don’t take 4%” or even “Don’t take 4.5%.” What I meant to say is don’t misrepresent it as scientifically verified “safe.”

I’d prefer to try to inform the investor of the range of result probabilities, as best we can show what they appear to be, as a basis for the investor's choice on the tradeoff of upside likelihoods v. downside risks.

And for that, I'd use Monte not historical-sequence, and graphit.

For example, with Monte applied to the Bengen/Wade data and allocations, the Bengen plan extended for 40 years (omitting fees and taxes) looks like this:

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/1 ... tr></table>[/img]

Good chances for terrific results. But on the other hand, risk of out-of-money 12% for 30 years, 23% for 40.

Below is a more realistic view, with 2% shaved off the portfolio’s “expected” return rate. That shaving is just a rough first stab at exploring potential effects of four reducers of the results: there may well be (a) taxes and/or (b) fees, and there are (c) the warnings of Mr. Bogle and others about potential low returns ahead and (d) the mathematical warning of Rodc that our inadequate historical data may have led us to overestimate the true probabilities hidden in that history. With the 2% shave from the portfolio's "expected" return rate, it looks like this::

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/v ... tr></table>[/img]

Risk of out-of-money 31% for 30 years, 48% for 40.

For the informed investor, it’s an informed tradeoff. The investor may well chose to go with 4.5%, hoping for the very good probabilities of doing very very well – but also informed that it’s not entirely “safe.”

++++++++

Getting back to what should be labeled “safe” and seen by investors more concerned with safe than upside prospects, here is a graph-from-Monte-Carlo view of the Bengen/Wade data and allocations with 2% shaved from expected return rates to cover all that stuff cited above, showing survival probabilities for a whole field of combinations of withdrawal rate and number of years:

[img]<table><tr><td><a%20href="https://picasaweb.google.com/lh/photo/U ... tr></table>[/img]

It starts with $1 million, and the curves represent annual withdrawals as listed in the box in the upper right, top down. The lowest line is $45,000 drawn, 4.5% (inflation-adjusted), up to the top curve representing $25,000 or 2.5%.

It’s this kind of analysis that leads me to look askance at the bandying about of any withdrawal rate above 3% as scientifically proven “safe.”

And it's likely, I hope, to help spark some of us to seek and demand more lifetime-SPIA purveyors to offer them with inflation adjustment. They can offer to roughly double that 3%.

Dick Purcell
Last edited by Dick Purcell on Tue Jun 21, 2011 1:58 pm, edited 2 times in total.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

DRiP Guy wrote:
Leesbro63 wrote:What is the current PE10?
http://www.multpl.com/
Not full blown "scary" but certainly "uncomfortable".
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

DRiP Guy wrote: What does 'drop out' & 'quit' mean in this context?
In the direct sense it means that people give up. Like the middle class is doing with regard to funding college for their kids. The goalpost keeps moving farther and farther away, so they just throw their hands up and say "I can't do it". Same for retirement saving. In the indirect sense it's just another symptom of the squeezeflation that is killing the American Dream for the middle class.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 4:54 pm

Post by DRiP Guy »

Leesbro63 wrote:
Yipee-Ki-O wrote: Yes, that would be the one and only Robert Michael "Hocus" Bennett.
Dare I ask whatever happened to him?
Google is your friend.

I assume your interest is natural and reasonable. However, if you know anything of this board's history, then you know it was created specifically to escape his incessant trolling at the previous home of the Bogleheads. Not an opinion, a well-documented fact.

For that reason, I suspect you will find that continuing to bring up his name or his activities will not find much favor here.

I, of course, speak only for myself, but again, since this board's inception, it has (at least partly for that reason I'm sure) had an enormously and refreshingly low tolerance for trolls and spammers.

That is at least in part what makes it such a pleasant place to come, discuss, share, and learn.

I hope we keep it that way.
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

I didn't mean to stir the pot. And I never realized how bad it was. But I did know he was an issue and since his name came up, I just wondered. Sorry I asked!!
Yipee-Ki-O
Posts: 167
Joined: Sun May 24, 2009 12:42 pm

Post by Yipee-Ki-O »

Leesbro63 wrote:I didn't mean to stir the pot. And I never realized how bad it was. But I did know he was an issue and since his name came up, I just wondered. Sorry I asked!!
Not to worry! If you've Googled him I believe you can understand why he's been "uninvited" by Bogleheads and so many other personal finance discussion boards and blogs over the years. Alas, he was just unable to follow this bit of advice Scott Burns gave him years ago regarding Mr. Bennett's oft-expressed beliefs regarding SWR's; "You go about it in a manner that is catastrophically unproductive by adding missionary zeal that inflates your importance and demeans others. The whole idea that there is a new school of Safe Withdrawal Rates reeks of personal aggrandizement."
Leesbro63
Posts: 6813
Joined: Mon Nov 08, 2010 4:36 pm

Post by Leesbro63 »

I remember him from MOTLEY FOOL long ago before I found BOGLEHEADS. YOI
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

Bongleur wrote:Wade, re the June 12 tabular info you posted...

Suppose you add a 4th column, showing how many more years beyond 30 are supported by the same 4.5% _real_ draw rate, based on each cohort's remaining portfolio amount? It ranges from none to maybe 120 _more_ years, in the case where the ending _real_ value of the portfolio is more than 4x the starting portfolio ??? Something is fishy here... am I reading it correctly?

I wish 30 years was not the default standard timespan for these studies; I think lifespans are such that all work should use 40 years as the amount of time you need to plan for the money to last.

DEF Cohort: each retirement starting year is a cohort.
Bongleur, I already had a table sort of like this made up, by your post inspired me to actually do something with it. I've made a blog post now, which includes a table allowing you to track the evolution of real wealth for up to 60 years (less than this for retirees since 1950) for each retiree. This table is rather massive and will not look nice if shrunk to 700 pixels. Please see here:

http://wpfau.blogspot.com/2011/06/retir ... tover.html

Sometimes, real wealth can grow very high over time. It is when the actual MWR for that retiree is much higher than the 4.5% used in this case. It also depends on the sequence of returns, not just the MWR. But you are indeed reading that correctly.
User avatar
Topic Author
wade
Posts: 544
Joined: Fri Sep 17, 2010 12:38 am
Location: Main Line
Contact:

Post by wade »

1210sda wrote:Wade's research paints a tough picture . For 2010 retirees, he cites a MWR of 1.8%. That's 55% below the popular 4%.

Wade states his hope that his predictions for recent retirees are too low. Me too.

1210

P.S. back to the drawing board
Yes, I'm predicting 1.8% for the 2010 retiree. However, I've been making efforts to explain better why this is hopefully too low. Basically, the situation we are dealing with these days has not been experienced in previous US history. This makes it hard for the regression to make proper predictions, as it must predict outside the range of historical experience. The opposite of this happened in the period before 1920 when valuations were very low and my model predicted higher sustainable withdrawal rates than what happened in reality. It is reasonable to think the opposite could happen now: valuations were high and the model predicts lower withdrawal rates than what will happen in reality.

There is a world of difference between 1.8% and 4% though. DRiP Guy and I discussed this in December when I finished an earlier draft of that paper. If I was actually approaching retirement, I would probably try to aim for about 3% (assuming I could afford it). However, I would not view that as ultra-safe by any means, and I would be monitoring the evolution of my current withdrawal rate after retirement.
Locked