4% safe withdrawal rule: does it work for other countries?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

4% safe withdrawal rule: does it work for other countries?

Post by LH »

I was reading the 4 % SWR thread on:

http://www.early-retirement.org/forums/ ... 19234.html

Which is real nice thread.

This SWR rate is calculated using the US results. I was wondering what the equivalent SWR would be in say Germany, Russia, Britain, Spain, Japan, using their stock and bond market returns?

Is our 4 percent SWR just an anomaly, a result of a "lucky" (used very loosely, but we were lucky we had the ocean between us, just like britain was lucky to a lesser extent they had the channel) country that had the good fortune to not be devasted in WWII for example. A country that has been politically stable in a relative sense to say, germany, or russia?

Will there be reversion to the mean for the US as a country in our lifetimes?

Perhaps goofy question in some regard, but if one simply extends reasoning of SWR calculations, reversion to the mean amongst countries (all roads no longer lead to rome, britannia no longer rules, etc), one easily runs into some dilemmas.

One has to go of course with something, and to consider that the past experience of the US will continue in terms of excellant stock market performance relative to other countries is reasonable, but then again it may just be recency bias to? Why should we not regress to the mean of countries?

I think it might have been 4 pillars that stated that back in 1900, no one would have predicted the US would become dominant, the US was an emerging market then........ (very loose recollection, may be off).

So might not a more reasonable approach to 30 year planning include a more inclusive weighting of country risk, given that the US might have just "lucked out" in some sense, and be an outlier?

sorry for the rough nature of the post. Kinda a hazy idea. But say I lived in Germany, what would people be telling me my SWR rate is based on the german experience, which yeah, does include losing two wars, but conversly, the US experience includes winning two wars which were not fought on its territory, a real nice advantage? Ditto japan. Seems SWR might be vastly different, if calculated. Was the US destined to win the wars ex ante? It just seems like sure we won, but what if we had lost, lets make random stuff up and say china joined japan, and flooded our country with soldiers after thier pearl harbour attack, and our country was devasted...we lost initially, the west was occupied and destroyed in sherman like fashion. (yeah historically goofy?) what would our SWR be then?

So our SWR rate is actually dependant on the fact that china and japan did not join forces, japan using its industrial might to ship over waves and waves of chinese against us or something like that,ridiculous I know historically..... But just an attempt to bring up for discussion things that our nice SWR depends on I think.

Using the US results seems to be a best case scenerio, which I do not neccessarily see why they will be repeated over the next 30 years. Maybe we will fall down to middle of the pack results in terms of countries stock market results, with a different resultant SWR.

To me that SWR rate actually seems to assume a whole lot, that may or may not be true going forward?
User avatar
mudfud
Posts: 1235
Joined: Tue Feb 20, 2007 3:34 pm

Re: 4% safe withdrawal rule: does it work for other countrie

Post by mudfud »

LH wrote:

To me that SWR rate actually seems to assume a whole lot, that may or may not be true going forward?
LH,

Yes, basically you are correct. As you imply, there is nothing magical about 4%; the SWR will be different in other countries. Also, the SWR calculations don't really take into consideration conditional probability (so this is somewhat similar [but not identical] to the Gambler's fallacy). Finally it's based on the assumption that the future will resemble the past--this was the SWR, but you don't know if this will be the SWR.

So there will never be a precise answer. The key I think is to use adaptive rules, and thereby (indirectly) implement conditional probabilities.

Mud
"Are you sure you have tested an a priori hypothesis?" | | Image
User avatar
ken250
Posts: 1894
Joined: Mon Feb 26, 2007 7:50 pm
Location: US-101

Post by ken250 »

LH,

Why not calculate it yourself?

Good Luck.
bpp
Posts: 2017
Joined: Mon Feb 26, 2007 11:35 am
Location: Japan

Post by bpp »

LH,

4% would not have worked in Japan, assuming the canonical 100% domestic, 75/25 stock/bond portfolio. A Japan-based investor who retired starting 1 Jan. 1990 would have already gone broke by now.

I don't have data further back than that, but for reference, here are some numbers on how a Japanese retiree would stand as of 1 Jan 2007. I assume 4% or 3% withdrawals in the first year (1990), increase the withdrawals for inflation (Japanese CPI), and show what percentage of the portfolio would be being withdrawn in 2007:

Code: Select all

Column 1:  Domestic stocks/Foreign stocks/Domestic bonds/Foreign bonds
Column 2:  2007 withdrawal rate

4% initial withdrawal rate:
50/00/50/00     29%
25/25/50/00      6.7%   
25/25/25/25      6.2%

3% initial withdrawal rate:
50/00/50/00      8.6%
25/25/50/00      3.5%   
25/25/25/25      3.3%
Notes:
  • 1) All portfolios are 50/50 stock/bond
    2) A zero expense ratio was assumed, so real-world performance would be a bit worse.
    3) "Foreign bonds" = US commercial paper.
Comments:
  • 1) A pure domestic portfolio was a disaster.
    2) International diversification of stocks made a big difference.
    3) International diversification of bonds made a small improvement, but not really make-or-break.
    4) With international diversification of at least stocks:
    • a) 4% looks like it may have a reasonable chance to survive the standard 30 years for SWR study purposes. (Just a guess.)
      b) 3% looks like it may survive much longer.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Post by LH »

The key I think is to use adaptive rules, and thereby (indirectly) implement conditional probabilities
Yeah, thats a more elegant way of describing the lines I am thinking along.
ken250 wrote:Why not calculate it yourself?

Good Luck.
I would need lots of luck, I cannot even calculate SWR for USA : )

Even assuming I knew how to calculate US SWR, japans would be tough for me, since our data,say on yahoo or whatever, is in us dollars for japan, and what one needs to have(i think) is stock market and bond data delineated directly in japan dollars, otherwise currency effects may wreck the calculations? Someone here once commented that a stock market graph looks very different if its delineated in japan dollars than us dollars. It was around the time of the dow elevation articles saying its an artifact of us currency, and if one was a forieng investor in the dow, that one really did not gain much due to currency, or something.

Bpp calculation shows japan would likely blowup since they are withdrawing 29 percent today under a 4 percent regimen starting in 1990.

That is something to be seriously considered I think?

Japan since 1990 has not really undergone any huge horrible shocks, it is a modern country, was even purported to be the predominate country going forward here in our own country(japan inc and such) during some of that period. Yet a simple 4 percent withdrawal rate fails there from 1990 to the present(likely fails, hard to imagine having to take out 29 percent and recovering from that, but possible I guees?)....

I would not have guessed that failure at all. Thanks for the japan calculation. Makes me wonder even more where the US falls in the continuium of modern, first world countries in terms of swr?

Maybe a more realistic SWR is 3 percent to include for japan like downturns that could happen here in the US, just as they happened in japan? Or maybe other countries like sweden would have SWR like 5 percent or something, and it would average out to 4 percent still?

Thanks for the replies,

LH
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

The SWR rule of thumb

Post by bobcat2 »

The SWR is a static rule of thumb that is a poor retirement strategy. Here are excerpts from a recent paper by Bill Sharpe on the efficacy of the 4% rule.
An efficient retirement strategy must be totally invested in the risk-free asset to provide constant spending in every future state. However, the generic 4% rule couples a risky, constant-mix investment strategy with a riskless, constant spending rule. There is a fundamental mismatch between its strategies, and as a result it is inefficient….

The 4% rule does not generate a revealed utility because the investment and spending rules do not correspond to an efficient retirement strategy. Retirees interested in fixed retirement spending should invest in the risk-free asset. Anyone who chooses to invest in the market should be prepared for more volatile spending. Either can adopt an efficient strategy. However, a retiree who plans to spend a fixed amount each period, while investing some or all funds in the market, faces a very uncertain future. Markets could perform well, and his wealth would far exceed the amount needed to fund his desired spending, or they could perform poorly, and his entire spending plan would collapse….

Unfortunately, the 4% rule represents a fundamental mismatch between a riskless spending rule and a risky investment rule. This mismatch renders the 4% rule inconsistent with expected utility maximization. Either the spending or the investment rule can be a part of an efficient strategy, but together they create either large surpluses or result in a failed spending plan.
Here's a link to the paper. Click on "Efficient Retirement Financial Strategies". See pages 20-23 of the paper for more on SWR's.
Link: http://www.stanford.edu/~wfsharpe/retecon/index.html

Sensible retirement strategies focus on keeping the individual's standard of living as smooth and as high as possible during retirement. This calls for a flexible withdrawal rate that varies with market returns. If you don't want your withdrawals and standard of living to vary much over time, invest in a less risky portfolio. Otherwise prepare for a bumpy ride.

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
rmark1
Posts: 359
Joined: Sat Jun 16, 2007 7:43 am

Post by rmark1 »

LH, check out the book ‘Triumph of the Optimists” by Dimson, Marsh, and Staunton. They show the returns and volatility for 15 or 16 developed world countries, including the US, 1900 – 2000. I ran a quick and dirty test using the Moneychimp.com monte carlo demonstrator, setting up a beginning $1000000 portfolio with a $40000 withdrawal. The US data gave an about 85% success rate (consistent with most US studies), with the lowest being Italy at about 13%, suggesting the US is the anomaly not the rule.

I’ve finally concluded that most recent withdrawal studies are just silliness, as authors churn through small data sets pretending to find complicated withdrawal rules. Add to that the consistently reported fact that most mutual fund investors receive below market returns (costs, dollar weighting) makes most study results optimistic. Hebeler seems to be the only one using realistic costs with past US history.

As a result I’ve adopted a simple plan –
Your best guess of future return is current valuation, life expectancy for withdrawal rate, asset value for future spending, withdrawal net of pension cash flows for bond allocation.

Divide after tax fixed pensions and mixed portfolios annually over IRS life expectancy,spending the lesser of the new divisions or the previous yearly divisions plus inflation.Portfolio balance unused for spending is a reserve for large purchases and emergencies.

Conveniently this gives a 4% withdrawal at a 25 year life expectancy (consistent with US studies), its inflation adjusted IF investments do well, adjusts to value as part of the plan instead of making it up as I go along, incorporates fixed pensions into withdrawals, and I even use it to roughly set my allocation. Overly conservative for past US history but maybe a smooth decline into genteel poverty in other histories. Unlike many authors, my crystal ball is a little cloudy.

My opinion, which may be worth what you are paying for it.
Badinvestor
Posts: 801
Joined: Thu Mar 01, 2007 1:10 am

Post by Badinvestor »

The 4% rule works better in foreign countries because they don't have the health care inflation that America does, they don't face issues like Medicare going bankrupt, and so expenses of retirees in those countries can probably be expected to rise more along the lines of the respective CPI.

If you are in a major developed country and you think US stocks will give you an edge over that country's stocks based on the US stocks' track record, you can always overweight US stocks. You may have some difficulty with exchange rates, though.

The Japanese case is not really an argument against 4% because the same thing can happen anywhere. 4% is not guaranteed and bad stock market performance will bust it. But 4% works better where you have more control over your expenses in retirement, as is the case in the rest of the developed world.
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

rmark1's plan

Post by bobcat2 »

For the most part I find rmark's retirement plan to be well thought out. It appears to be a seat of the pants form of consumption smoothing.

I do have one big objection, however. He bases the retirement plan on life expectancy, which means the retiree has a 50% probability of outliving the plan. This is definitely not a good idea. A more prudent approach is to base the plan on maximum life expectancy. For most people under age 85 a reasonable estimate of maximum life expectancy is age 100.

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
sscritic
Posts: 21853
Joined: Thu Sep 06, 2007 8:36 am

Re: rmark1's plan

Post by sscritic »

bobcat2 wrote:For the most part I find rmark's retirement plan to be well thought out. It appears to be a seat of the pants form of consumption smoothing.

I do have one big objection, however. He bases the retirement plan on life expectancy, which means the retiree has a 50% probability of outliving the plan. This is definitely not a good idea. A more prudent approach is to base the plan on maximum life expectancy. For most people under age 85 a reasonable estimate of maximum life expectancy is age 100.

Bob K
rmark wrote:
Divide after tax fixed pensions and mixed portfolios annually over IRS life expectancy
My reading of rmark is that you do the equivalent of RMDs, recalculating your life expectency each year using the Universal life table, i.e., even at age 100, your life expectancy is 6.3 years. It doesn't run out for anybody. Even at age 115, you only spend half your money each year. It may get smaller by the time you are 130, but it never runs out.
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Post by bobcat2 »

If you do this you are running the risk of greatly reducing your standard of living once you outlive your life expectancy. Better to base the plan on maximum life expectancy from the beginning.

That's why I added age 85. Once you reach age 85 you would want to bump up age of death beyond 100. You never want to base the plan from now forward on life expectancy. You will always face a 50% probability of outliving the plan if you do so. Even at age 100 you would base the plan on an age beyond your life expectancy at that point.

The goal should be to keep your standard of living relatively smooth over your retirement. The goal should not be simply avoiding running out of money except for social security. Recalculating life expectancy every year will reduce your living standard every year you survive. This could be a very significant reduction in living standard if you retire at age 61 and are still quite alive at age 86.

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
lazyday
Posts: 3849
Joined: Wed Mar 14, 2007 10:27 pm

Post by lazyday »

LH wrote:I would not have guessed that failure at all.
Sorry if sounds cocky, but I would have. In 1991 or 1992 when I was choosing mutual funds, Japan's market was at crazy P/E and other valuation multiples. I refused to hold any fund which had more than a tiny holding in Japan. (But unfortunately for me, I strongly favored foreign holdings to U.S. at a time when the U.S. market was going to outperform.)

It seemed clear to me at the time that Japan's market was so overvalued, that returns very likely would be quite poor.

I think it is important to not only look at past returns or returns sequences (which is the main way I know of to come up with the 4% number) but to also consider current prices and where future returns should come from. Not that it's easy to do so, but it could have saved you in Japan '89, U.S. early 2000, and for sure from internet stocks.

For example, one might start with earnings yield. I believe it generally needs to be adjusted downward for a few reasons, but if even before those adjustments, the earnings yield is below 4%, then I wouldn't think 4% is a safe rate going forward. (Unless perhaps recent earnings are very depressed. They weren't in U.S. in March 2000; I don't know about Japan '89.)
rmark1
Posts: 359
Joined: Sat Jun 16, 2007 7:43 am

Post by rmark1 »

“The 4% rule works better in foreign countries because they don't have the health care inflation that America does, they don't face issues like Medicare going bankrupt, and so expenses of retirees in those countries can probably be expected to rise more along the lines of the respective CPI.”

See my first post, I ran a short test using a 60/40 market weight portfolio similar to what most withdrawal studies use against the data in the book, and the 4% had low success rates likely due to the lower returns and higher volatility in most foreign markets. Most other developed world countries have some form of government social/medical insurance, and face the same demographic problems the US faces, some with higher debt/GDP loads than the US. There is no magic overseas.

In essence, I’m setting upper and lower withdrawal bounds using a combination of life expectancy and asset value. Life expectancy moves out as you age – you’re expected to live a longer shorter time. Eventually the increasing withdrawal rate results in a late life spend down. It is a mickey mouse form consumption smoothing.

I just use stock earnings yield – costs and bond yield – inflation - costs as a best guesses of future return.
jar2574
Posts: 258
Joined: Mon Mar 19, 2007 2:19 pm

Re: rmark1's plan

Post by jar2574 »

bobcat2 wrote:For the most part I find rmark's retirement plan to be well thought out. It appears to be a seat of the pants form of consumption smoothing.

I do have one big objection, however. He bases the retirement plan on life expectancy, which means the retiree has a 50% probability of outliving the plan. This is definitely not a good idea. A more prudent approach is to base the plan on maximum life expectancy. For most people under age 85 a reasonable estimate of maximum life expectancy is age 100.

Bob K
Great points. I like the idea of recalculating life expectancy each year.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Post by LH »

rmark1 wrote:“The 4% rule works better in foreign countries because they don't have the health care inflation that America does, they don't face issues like Medicare going bankrupt, and so expenses of retirees in those countries can probably be expected to rise more along the lines of the respective CPI.”

See my first post, I ran a short test using a 60/40 market weight portfolio similar to what most withdrawal studies use against the data in the book, and the 4% had low success rates likely due to the lower returns and higher volatility in most foreign markets. Most other developed world countries have some form of government social/medical insurance, and face the same demographic problems the US faces, some with higher debt/GDP loads than the US. There is no magic overseas.

In essence, I’m setting upper and lower withdrawal bounds using a combination of life expectancy and asset value. Life expectancy moves out as you age – you’re expected to live a longer shorter time. Eventually the increasing withdrawal rate results in a late life spend down. It is a mickey mouse form consumption smoothing.

I just use stock earnings yield – costs and bond yield – inflation - costs as a best guesses of future return.
I would agree here entirely, I almost replied along the same lines, but really I know very little about overseas, my strong gestalt is they have more socialism than we do, more problems with programs thats are simply not sustainable, even in the short term, than we do, and all seem to be moving towards more free market, less regulation, or at least seem to acknowledge they should be. They have to compete against us, compete against asia, etc. But thats just my opinion really, I cannot quote many facts and such.
Badinvestor
Posts: 801
Joined: Thu Mar 01, 2007 1:10 am

Post by Badinvestor »

rmark1 wrote:See my first post, I ran a short test using a 60/40 market weight portfolio similar to what most withdrawal studies use against the data in the book, and the 4% had low success rates likely due to the lower returns and higher volatility in most foreign markets.
What makes you think that people living in foreign countries are going to limit their equity investments to their own country instead of diversifying globally?
Most other developed world countries have some form of government social/medical insurance, and face the same demographic problems the US faces, some with higher debt/GDP loads than the US. There is no magic overseas.
The magic is that health care prices are lower and are increasing more slowly abroad. In addition, providers there are not as aggressive as providers are here about increasing utilization to keep their revenues rising at a desired rate. The providers abroad are often state entities or state employees whose compensation is not based on a fee for service model.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Post by LH »

lazyday wrote:
LH wrote:I would not have guessed that failure at all.
Sorry if sounds cocky, but I would have. In 1991 or 1992 when I was choosing mutual funds, Japan's market was at crazy P/E and other valuation multiples. I refused to hold any fund which had more than a tiny holding in Japan. (But unfortunately for me, I strongly favored foreign holdings to U.S. at a time when the U.S. market was going to outperform.)

It seemed clear to me at the time that Japan's market was so overvalued, that returns very likely would be quite poor.

I think it is important to not only look at past returns or returns sequences (which is the main way I know of to come up with the 4% number) but to also consider current prices and where future returns should come from. Not that it's easy to do so, but it could have saved you in Japan '89, U.S. early 2000, and for sure from internet stocks.

For example, one might start with earnings yield. I believe it generally needs to be adjusted downward for a few reasons, but if even before those adjustments, the earnings yield is below 4%, then I wouldn't think 4% is a safe rate going forward. (Unless perhaps recent earnings are very depressed. They weren't in U.S. in March 2000; I don't know about Japan '89.)
In terms of being cocky, I do not think it applies to ex post statements personally.

Munckin man is cocky, and refreshing in a way. Ex post is trivial.

Stating you knew something ex post does not mean much of anything in terms of what percentage of things you thought you knew for certain actually turn out to be true. The human mind is a horrible objective tracker of this, as I think behavoiral science has shown.... The things you KNEW would happen, that do not occur, you tend to forget. Just like I easily remember the stocks mcd, eslr, nvidia I bought and sold. (I smile and feel real pleasure when I write those out)... ah the dopamine surge : ) I KNEW IT!

; )

So remembering you knew something for certain, doesnt mean it had any meaningful predictive value or worth, just because it came true. You would have to remember and list out all the things you knew to be true, and see what happened percentagewise for all of them. Where humans actually have a nice record of acting on thier valuations for all to see, aka managed mutual funds, the results are not pretty versus mr markets indexes. So while your valuation timing may have saved you money in that instance, long haul, you likely have lost more money using that technigue than gained, aka mutual funds.


PS dopamine surge reference is from your money and your brain. Sorry for the jargon. Its basically when you gamble, and get it right, your brain releases the compound dopamines, and it gives you pleasure, and reinforces that experience, makes it more memorable, and makes you more likely to do that behavoir again.

I still have to get off the crack of individual stocks myself. I probably will always play with em some, at least I buy em for years, and not make short term bets, so I get a nice long roulette ball rolling for years....... But expectation is I will lose versus index long term doing that, but ah man, I hope I will be the exception! : )


LH
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Post by LH »

Thanks to bobcat for expanding my horizons and making me consider relearning caculus and such. Anyway, I kinda scimmed the paper, since I cannot do calculus really, it is still kinda beneficial and helpful I think. A nice summary of the 4 percent rule, then their critigue of it, which descriptively, makes little sense to me, but maybe the math makes up for it.

http://www.stanford.edu/~wfsharpe/retecon/ERFS.pdf
The 4% Rule
Many recent articles in the financial planning literature have attempted to answer the question: “How much can a retiree safely spend from his portfolio without risking running out of money?” Bengen (1994) examined historical asset returns to determine a constant spending level that would have had a low probability of failure. He concluded that a real value equal to approximately 4% of initial wealth could be spent every year, assuming that funds were invested with a constant percentage in equities within a range of 50% to 75%. Cooley, Hubbard, and Walz (1998) used a similar approach and found that a 4% spending rule with inflation increases had a high degree of success assuming historical returns and at least a 50% equity allocation. Later, Pye (2000) concluded that with a 100% allocation to equities, the four percent rule would be safe enough if equity returns were log-normally distributed with a mean return of 8% and a standard deviation of return equal to 18%.4 Based on this research, there is a growing consensus that newly retired individuals with funding horizons of thirty to forty years can safely set their withdrawal amount to 4% of initial assets and increase spending annually to keep pace with inflation. This is the foundation for the now common 4% rule of thumb for retirement spending.

An efficient retirement strategy must be totally invested in the risk-free asset to provide constant spending in every future state.5 However, the generic 4% rule couples a risky, constant-mix investment strategy with a riskless, constant spending rule. There is a fundamental mismatch between its strategies, and as a result it is inefficient. The following simple example illustrates these points. Consider a retiree who, whether the
market goes up or down, wants to spend only $1 next year. He can achieve this goal byinvesting 1/Rf dollars in the risk-free asset. On the other hand, if he uses the market asset, he must increase his investment to 1/Rd dollars, so that if the market goes down, the investment pays the required $1. However, if the market goes up, the investment pays (Ru/Rd) dollars, and there is an unspent surplus. So, if our retiree truly requires just $1, then investing in the market is less efficient than investing in the risk-free asset because of the greater cost and the potential unspent surplus.
In terms of potential unspent surplus being not efficient, I dunno if I want efficiency per se.

I cannot follow the math, but commonsensewise(heh, citing commonsense shows my lack of understanding), this makes no sense.

It takes the downside of investing in the market, requiring MORE money to deal with the potential volatility, and states thats a bad thing, fine.

But then it cites "truly requires" and makes the corresponding potential upside benefit of the market, unspent surplus, a bad thing too.......

I dunno. I do not know the maths, but I doubt many human being fit thier scenerio of "truly requires" the set amount, such that the upside benefit of unspent gain, is not, well, an upside benefit, and not a negative?

Its things like this that makes me view, unjustly perhaps, this whole thing as being in complete flux, and positions being taken perhaps more for advancement of learning, with unrealistic assumptions, so as to move the field forward and such. "Efficient" seems pretty contrived, but I cannot truly follow it, as I cannot do calculus.

So, the potential upside, as well as the potential downside of risky investment are both downsides. Man, just makes no sense.

Ah well, so speaketh the peanut gallery. Like the guy who painted my house critiqing michaelangelo I guess : ) Maybe someone more versed could interpret for me?

Cheers,

LH
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Financial economics viewpoint of SWR's

Post by bobcat2 »

Let’s take two examples. Two guys retire at the same age with exactly the same amount of assets and the same 60/40 asset allocation. The difference is the one guy retires on 12/31/96 and the other retires on 12/31/99.

They both expect their portfolios to return an average real return of 5% per year. The first guy after three years has a portfolio much larger than he expected. It makes little sense for him to continue to have a withdrawal rate of real 4% every year based on the value of the portfolio when he retired. He can easily make his AA less risky, increase his reserves, and increase his withdrawal rate and thereby his living standard. Why should he simply sit there and pile up a huge amount of excess reserves well above what he considered necessary when he planned his retirement?

The second guy’s prospects are much dicier. After three years of retirement he has a portfolio much smaller than he expected. Unless reversion to the mean occurs relatively quickly, he faces the prospect of portfolio extinction years before he dies if he continues to withdraw 4% real every year based on the value of the portfolio when he retired. He needs to decrease the withdrawal amount and his living standard. Staying with the SWR is an extremely risky strategy in this case.

A SWR policy does not adjust to the size of the portfolio now. It is in this sense that picking a SWR is a static policy.

If you want a smooth retirement ride you need to have a low risk portfolio. This will provide a relatively smooth consumption path and picking a SWR won’t be too far off, although it will still not be an optimal policy.

If, on the other hand, you want a riskier portfolio with a higher expected consumption path, picking a SWR could easily result in either piling up large unnecessary reserves or portfolio extinction. In this situation a SWR strategy is far from optimal.

The SWR rule of thumb appears to be an effort by financial practitioners to make a risky retirement portfolio less risky. It doesn't work. Either the portfolio is risky or it isn't. You cannot make a risky retirement portfolio less risky by picking a static withdrawal rate.

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
rmark1
Posts: 359
Joined: Sat Jun 16, 2007 7:43 am

Post by rmark1 »

“What makes you think that people living in foreign countries are going to limit their equity investments to their own country instead of diversifying globally?”

Currently I think they would globally diversify. However the 4% of initial portfolio value+ annual inflation rule was derived solely from past US data, and I wanted an apples to apples test using past individual countries markets, not a test of broader diversification. There is a tendency to compare US to international as a whole, which I feel is more apples to oranges, and which really wasn’t available historically.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Re: Financial economics viewpoint of SWR's

Post by LH »

bobcat2 wrote:Let’s take two examples. Two guys retire at the same age with exactly the same amount of assets and the same 60/40 asset allocation. The difference is the one guy retires on 12/31/96 and the other retires on 12/31/99.

They both expect their portfolios to return an average real return of 5% per year. The first guy after three years has a portfolio much larger than he expected. It makes little sense for him to continue to have a withdrawal rate of real 4% every year based on the value of the portfolio when he retired. He can easily make his AA less risky, increase his reserves, and increase his withdrawal rate and thereby his living standard. Why should he simply sit there and pile up a huge amount of excess reserves well above what he considered necessary when he planned his retirement?

The second guy’s prospects are much dicier. After three years of retirement he has a portfolio much smaller than he expected. Unless reversion to the mean occurs relatively quickly, he faces the prospect of portfolio extinction years before he dies if he continues to withdraw 4% real every year based on the value of the portfolio when he retired. He needs to decrease the withdrawal amount and his living standard. Staying with the SWR is an extremely risky strategy in this case.

A SWR policy does not adjust to the size of the portfolio now. It is in this sense that picking a SWR is a static policy.

If you want a smooth retirement ride you need to have a low risk portfolio. This will provide a relatively smooth consumption path and picking a SWR won’t be too far off, although it will still not be an optimal policy.

If, on the other hand, you want a riskier portfolio with a higher expected consumption path, picking a SWR could easily result in either piling up large unnecessary reserves or portfolio extinction. In this situation a SWR strategy is far from optimal.

The SWR rule of thumb appears to be an effort by financial practitioners to make a risky retirement portfolio less risky. It doesn't work. Either the portfolio is risky or it isn't. You cannot make a risky retirement portfolio less risky by picking a static withdrawal rate.

Bob K
Piling up "excess" reserves simply decreases his risk I think?

If I have a 60/40 portfolio, and it increases 40 percent one year:

These "excess reserves" held in the same portfolio, simply decrease my risk of not meeting my retirement needs. The concept of excess reserves seems, I dunno, lets say VERY artificial as most people in real life would consider it to thier financial situation..... Excess reserves might qualify for bill gates. For the average retiree, that money simply reduces the risk of not meeting ones needs.

Are you really saying that in a 60/40 portfolio, where you calculate say 1 million is LIKELY to get you by, but you have RISK it will not, that having 1.4 million the next year is somehow BAD or neutral, and somehow "excess", and somehow does not LOWER your risk and somehow does not make you MORE LIKELY to get by in your 60/40 portfolio???

I mean, you know a lot more than me about this I grant, and I appreciate your time, you are very helpful, but I am so completely lost here, that on the one hand, when the portfolio goes down for the one guy, his risk goes up, his situation becomes "dicey" while the guy whose portfolio goes up, well, his risk is not less, its simply discounted as "excess" reserves?

I simply posit, that in a 60/40 portfolio, if my needs are likely to be met at 1 million, they are more likely to be met, ie failure risk is less, at 1.1, 1.2. 1.4 million or whatever in real dollars. The guy whose portfolio went up, has less risk, the guy whose portfolio went down has more risk.

Thanks for your help, I like trying to understand this stuff,

LH

PS my gestalt of some of whats going on here, is similiar to ignoring friction in physics. Its great for discovery of principles, hashing them out(thats what these guys are doing now, hashing out the stuff we will be taught 10 years or so from now?), and once figured out, great for teaching and learning too, But when applied to firing an artillery shell in real life, the equations where friction is ignored do not work. Is that a good analogy for any of this?
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Post by bobcat2 »

Let’s say you are about 60 years old and planning your retirement. A lot of your consumption (living standard) in retirement will depend on the performance of your portfolio.

If you choose a low risk portfolio you know about how much income (interest and capital) that portfolio will generate during retirement. So you also know about how much relatively smooth consumption you can afford during retirement. Choosing to have or not have a SWR won’t make much difference.

Alternatively, you can choose to invest in a riskier portfolio. That portfolio will support an expected consumption path higher than the low risk portfolio will support. There are, however, two alternatives. There is a chance that the portfolio will perform either very poorly or very well. That's the nature of a risky portfolio.

If we combine the risky portfolio with the fixed real withdrawal amount it won’t make much difference if we should get about the expected return. Now consider the other 2 cases. If the portfolio does poorly the fixed withdrawal amount can lead to portfolio extinction well before your death. Not good!

On the other hand, if the portfolio does extremely well with a fixed withdrawal amount, you simply pile up reserves you didn’t think you needed when planning your retirement. So you took on extra risk and your reward for taking on that risk is extra reserves. That’s not much of a payoff. So the downside to taking on extra risk is a very low standard of living in late retirement. The upside to taking on extra risk is simply piling up excess reserves. That seems like an inefficient strategy to me.

Bob K

PS - I thought the most important part of the Sharpe paper was the following quote:
Virtually all retirees have an explicit or implicit retirement spending and investment strategy. What is striking is the gulf that exists between how financial economists approach the problem of finding optimal retirement strategies and the rules of thumb typically utilized by financial advisors.
I agree. The gap today is at least as big as it was 30 years ago when financial economists were recommending index stock funds and building a portfolio based on asset allocation. The financial advisors of that era were instead recommending buying portfolios of individual stocks that would beat market returns or, alternatively, buying actively managed stock funds that would beat market averages. Today's financial advisors have picked up on what financial economists were recommending 30 years ago, but they don't appear to have picked up on hardly any of the economists' recommendations since then. The advisors have instead filled in the missing gaps of 30 years ago with rules of thumb that appear to do as much harm as good.

For an explicit criticism of another financial planning retirement 'rule of thumb' click on the following link and then click on the pdf file titled Rules of Dumb.
Link:http://www.esplanner.com/illustrations.php
Read the linked article from CBSMarketWatch and compare the retirement advice from AARP, Garrett, et al in the CBSMarketWatch article to the Kotlikoff advice. As a matter of fact the 'expert' advice from AARP, Ginsburg, Ballpark Estimate, and Fidelity in the article does not rise to the level of ludicrous. In each case their advice to reach the retirement goal is to starve between the ages of 45-65. They apparently don't see any problem with this advice. :cry:
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Post by LH »

bobcat2 wrote: On the other hand, if the portfolio does extremely well with a fixed withdrawal amount, you simply pile up reserves you didn’t think you needed when planning your retirement. So you took on extra risk and your reward for taking on that risk is extra reserves. That’s not much of a payoff. So the downside to taking on extra risk is a very low standard of living in late retirement. The upside to taking on extra risk is simply piling up excess reserves. That seems like an inefficient strategy to me.
thanks for the reply bobcat : )

The "excess reserves" are not excess though? They reduce the future risk of your portfolio running out. This is a real benefit?

If you have a 60/40 portfolio, that goes up 30 percent in value due to the fact that the stock portion jumped significantly in the first year of retirement, your risk of running out of money is quite simply decreased.

You rebalance the portfolio, and you have more bonds and stocks than you did before.... This is not excesss reserves, this is risk reduction in action.

If I have 1.3 million in a portfolio, thats less risky than having 1.1 million in a portfolio, given a fixed 4 percent rate based on the initial portfolio, plus the inflation rate.

Having lower money due to bad performance=more risk of running out is the opposite side of the same coin of having higher money due to good performance=less risk of running out. Its a function of portfolio size and withdrawal rate and future expectations.

Higher portfolio size for a given withdrawal rate = less risk. The concept of "excess reserves" not lowering the risk I will experience in retirement boggles my mind.....

Give me the 1.3 million 60/40 portfolio versus the 1.1 million 60/40 portfolio with the same withdrawal per year. I dunno what I am missing, but the 1.3 sure seems to have less risk of running out.

I dunno if there is some sort of semantics going on? Perhaps we are talking past each other?

To me:

Risk==I run out of money early.

The concept of "excess reserves" ie more cash, not lowering my risk of running out of cash I cannot grasp. More cash=less risk of running out of cash to my simple mind. The reward is more cash, less risk? Argh. From my perhaps wrong viewpoint, its like you are saying black is white, and I am just repeating white is white over and over heheheheh.

Thanks for your posts, I like reading them. : )

LH

PS the whole 4 percent rule is predicated on the amount of cash available yielding a 4 percent safe withdrawal rate. The amount of money in the portfolio fundamentally matters to the level of risk present in relationship to the withdrawal rate, in fact, the amount of cash present determines what the safe withdrawal rate is.


That 4 percent SWR still has risk of running out. Now a 5 percent withdrawal rate is riskier than a 4 percent, which is riskier than a 3 percent.... So basically, if you have "excess reserves" under say an original 5 percent regimen, you will convert it to a 4 percent SWR regimen with less risk. If you start of with the 4 percent SWR, yet your portfolio jumps up enough, you are now having a 3 percent SWR, and less risk. You must be having some fundamental disagreement with this scenerio to begin with???? That the underlying premise of 3 percent safer than 4 percent safer than 10 percent withdrawal rate is untrue???

Fixing the amount withdrawn in real dollars at 4 percent of initial portfolio, but then having that portfolio value jump up enough in real dollars such that what you are withdrawing is actually only 3 percent... That must be a reduction in risk as I see it.... Or the whole 3 less risky than 4 less risky than 10 less than "90 percent "SWR"" must be somehow untrue? sorry for repetition. Interesting.

If the 4 percent SWR is truly "safe", I could then see your excess reserves argument, but in reality, 4 percent SWR is not perfectly safe is it? It just has a degree of safety, calculated from past us history. Maybe that is where we are talking past each other?

LH
sscritic
Posts: 21853
Joined: Thu Sep 06, 2007 8:36 am

Post by sscritic »

Here is my take on bobcat vs LH and how they are talking past each other.

bobcat wrote
Alternatively, you can choose to invest in a riskier portfolio. That portfolio will support an expected consumption path higher than the low risk portfolio will support.
In other words, you deliberately choose a riskier portfolio because you want to support a higher SWR. For example, one portfolio is predicted to support a 3% SWR. Another riskier portfolio is predicted to support a 4% SWR. By choosing the riskier portfolio, you get a riskier portfolio, i.e., a greater chance of running out.

LH assumes you are trying to support a fixed SWR and looking at the effects of using a riskier portfolio to support that fixed SWR.

Also, there is a difference of utility functions in play. Bobcat puts a very great value on not running out and a small value on having more than enough. LH thinks that having more than enough is a very valuable thing since he worries a lot and has trouble sleeping when he just has enough.

If I have misinterpret both authors, it won't be the first time. :D My apologies in advance to both of them for my errors.
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Problems with static withdrawal rate rules - Redux :-)

Post by bobcat2 »

Rule of thumb static retirement withdrawal rates

The main problem, as is often the case with these static and arbitrary rules of thumb, is the focus is on the wrong objective function. In personal finance the objective should be on keeping consumption as high and smooth as possible, not literally running out of money. That’s the basic problem with an arbitrary constant (static) retirement withdrawal rate.

If the retirement portfolio does extremely poorly during the first few years of retirement, a reasonable person (at a minimum) cuts her consumption target path over the rest of her life. Think of this as a simple TVM problem where PV is now significantly smaller and so the solution path of annual payments (portfolio withdrawals) is now lower. She also has many other options to consider, including going back to work and/or taking out a reverse mortgage. She does not say to herself, “Reversion to the mean is nearly in the bag, so I’ll keep using the constant withdrawal rate premised on the size of the much larger portfolio I ‘expected’ to have at this point in time.”

If, on the other hand, the retirement portfolio does extremely well during the first few years of retirement, a reasonable person can raise her consumption path, lower the risky percentage of her portfolio, purchase additional real life annuities, and increase her reserve fund. Why would she simply dump unexpected high returns into a reserve fund?

The only time any static withdrawal rate makes sense is when the retirement portfolio itself is extremely safe and can be expected to provide a flow of predictable smooth returns. For example, a retirement portfolio which consists mainly of laddered TIPS.

If you want to lower the risk to your retirement living standard then lower the risk in your retirement portfolio. Arbitrary static retirement portfolio withdrawal rates (including those based on MC analysis) give the illusion, but not the substance, of retirement income security.

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
User avatar
czeckers
Posts: 1082
Joined: Thu May 17, 2007 3:49 pm
Location: USA

Post by czeckers »

This is a good discussion, thanks!

The question of whether or not having a surplus is an interesting one. On one hand, LH is correct in that having more than expected obviously is not a bad thing in the sense that it either allows one to reduce the portfolio risk to maintain the same SWR or to increase the SWR amount proportionately while maintaining same risk.

However, it can be bad in the sense that it states in retrospect, that you took on more risk than you had to. By taking on the additional risk (for which you happened to be rewarded), you actually had a higher chance of running out of cash.

Unfortunatley you can only find out how things turned out after the fact, and not ex ante. Consequently, it seems prudent to set up your portfolio and withdrawal stragegy in such a way as to take on the least amount of risk to maintain a sustainable withdrawal rate. I think this is the bottom line. If you know that your cash-flow needs will be x. Ideally your portfolio would be all TIPS with yield matching x. Since few investors have enough assets to do this, it comes down to the trade-off between risk and reward.

-K

Conse
The Espresso portfolio: | | 20% US TSM, 20% Small Value, 10% US REIT, 10% Dev Int'l, 10% EM, 10% Commodities, 20% Inter-term US Treas | | "A journey of a thousand miles begins with a single step."
ataloss
Posts: 887
Joined: Tue Feb 20, 2007 2:24 pm

Post by ataloss »


The 4% rule does not generate a revealed utility because the investment and spending rules do not correspond to an efficient retirement strategy. Retirees interested in fixed retirement spending should invest in the risk-free asset. Anyone who chooses to invest in the market should be prepared for more volatile spending.





And allowing for volatility of spending should help the safe withdrawal rate- i.e. reduction of spending or at least forgoing inflation related increases increased portfolio survival in a review I saw AAII journal as I recall.

Sometimes people post as if future investment returns are the only variable and spending levels are certain but this isn't likely. A good friend of mine from northern Virginia enjoyed a tripling of his home value over a little more than a decade. He enjoyed the increases in taxes less and needs to sell but may find it difficult in the current market.
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 3:54 pm

Post by DRiP Guy »

ataloss wrote: A good friend of mine from northern Virginia enjoyed a tripling of his home value over a little more than a decade. He enjoyed the increases in taxes less and needs to sell but may find it difficult in the current market.

:lol: :lol: :lol:
User avatar
Topic Author
LH
Posts: 5490
Joined: Wed Mar 14, 2007 2:54 am

Re: Problems with static withdrawal rate rules - Redux :-)

Post by LH »

bobcat2 wrote:If, on the other hand, the retirement portfolio does extremely well during the first few years of retirement, a reasonable person can raise her consumption path, lower the risky percentage of her portfolio, purchase additional real life annuities, and increase her reserve fund. Why would she simply dump unexpected high returns into a reserve fund?

Bob K
Who is saying she would "simply dump unexpected high returns into a reserve fund" under the SWR scenerio?

I think I am envisioning a SWR as it would be practically applied by a boglehead, whereas you are thinking of a form of SWR as perhaps discussed in academic papers. We are talking past each other here.

I do not understand the "excess reserves" and the "dumping" terms.

Yes, the person could do all of those actions you list with the "excess reserves", and I expect they may (except perhaps early on they would not spend much more, the market gives and takes away). The "reserve fund" option, in this case, is the 60/40 portfolio, this is not "dumping" anything anywhere. One would be "dumping" money into stocks, and "dumping" money into bonds. One could choose to shift to more bonds, or add more annuities with this "excess" money too. Annuities, long term care plans, etc, are not precluded by someone using a SWR plan how I envision it.

Regardless of whether the "excess" money is wisely put into annuities, or "dumped" into bonds, that "excess" money still lowers risk.

I do not think I have enough knowledge of how these things are formally characterized to see exactly where you are coming from here, and perhaps you have too much knowledge to see what level I am coming from : )

Perhaps more money does not equal less risk under certain conditions in papers. In the real world, I am fairly certain if I envision spending X dollars a year in real terms, if I have a 1 million portfolio in real dollars in year zero, and 1.3 million real dollars in year 1, in a 60/40 portfolio, I have an expectation of risk that has gone down. But perhaps not. In the real world, I may well decide to spend X plus 1/100X additional if things go really well, because, well, I perceive the risk to have gone down and so I can spend some more. To me, this is "obvious", but in the SWR as formally discussed in papers, perhaps this is actually wrong. Spending more is not allowed perhaps either on annuities or on a vacation.

But If I had a portfolio, under a "SWR" plan, I personally would have annuities as part of the plan certainly already, if the market tanked I would cut back my consumption, maybe even go back to work. If the market went up 30 percent, I would likely "dump" some of that "excess" money into bonds via banded rebalancing into the bond portion- which I would consider to have lowered my risk, under the simplistic more bonds=less risk under a given withdrawal rate. I may well buy more annuities with it too. The SWR is just a tool to see where one stands in my book. It is just a rule of thumb, I do not see it as precluding annuities, precluding spending extra, or cutting back at all. I do not think most bogleheads view it as such, or practice it as such? I maybe wrong, I certainly speak for no one but myself.

I think the academic papers make stipulations that are not realistic but highly useful, much like the ignore friction in physics. I doubt many people who use a SWR in practice follow whatever the formal stipulations are in the SWR papers, they pay commissions, fees, taxes, modify consumption, go back to work(ie modifiy thier job production curve heheh), have more classes than the simple 60/40 portfolio, use annuities, pensions, social security, long term care plans, etc.....

Anyway, I think you are talking over my head, and I am talking under yours : ) I enjoy your posts a lot.

Thanks for your help bobcat2,

LH
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

One More Time - Why SWR is a bad rule

Post by bobcat2 »

Here’s my understanding of the SWR rule-of-thumb. At the beginning of retirement the retiree calculates a percentage of his portfolio to withdraw and spend that year. He then adjusts that amount for inflation each year during retirement.

Example – He retires at age 65 with a million dollar portfolio. Using the 4% SWR rule-of-thumb he withdraws $40,000 to spend the first year. He then adjusts the $40,000 for inflation each year during retirement. So he withdraws/spends a constant $40,000 per year real during retirement and thus he should avoid running out of money.

It appears to me that you want to use and/or improve the above SWR rule-of-thumb. I, on the other hand, do not want to use it or improve it. I think it’s a bad rule that should not be used by anyone.

To me it’s a bad idea because it tries to marry a constant real withdrawal/spending amount with volatile investment returns. So in many cases the rule will be abandoned because either (a) a retirement return sequence will be very poor (2000-2002) and you may run out of money, or (b) the retirement return sequence will be very good (1996-1999) and you will have a spending plan that is unduly low given your asset level. In either case the SWR rule should, and in all likelihood will, be abandoned. So what good is such an ineffectual rule?

Here’s one more way of seeing what’s wrong with this approach. The SWR methodology runs Monte Carlo simulations to determine the household's probability of running out of money. These simulations assume that households make no adjustments to their spending regardless of how well or how poorly their investments do. However, a consumption smoothing objective function dictates such adjustments and precludes running out of money, i.e. dropping consumption to zero. But it is the range of these living standard adjustments that households need to understand to assess their portfolio risk. Thus the SWR method not only asks the wrong question (How do I avoid running out of money?) instead of the right question (How do I keep my standard of living high and relatively smooth during retirement?). But it may also improperly encourages risk taking, because it focuses on the fact that risky investments, due to their higher expected mean return, may entail a lower chance of financial exhaustion, rather than focusing on the very real risk of either poor or bad risky portfolio returns.

So if you don’t use a SWR rule what can you do? There are several sophisticated strategies that could be adopted such as using a lockbox approach or dynamic programming methods, but a simple method would be to use a TVM amortization solution for our 65 year old retiree. Calculate the portfolio’s expected return. If the portfolio is less than 50% risky assets drop the expected return by one percent to be conservative. If the portfolio is more than 50% risky assets drop the expected return by as much as two percent in the analysis. Assume a maximum life age of say 100 in the analysis.

So at age 65 with a 60/40 portfolio and an expected portfolio real return of 4.0% the problem becomes:

pv = -$1,000,000
i = 2.6% (4.0% minus 1.4%)
fv = 0
t = 35
payment = ? (solution about $43,900)

The main difference between this and the SWR is you redo this calculation every year. (It takes about a minute.) You need to do this every year to keep your consumption (withdrawal amount) path consistent with the current size and current expected return of your volatile portfolio. If you want a smoother consumption path, you will need to adopt a more conservative portfolio with a lower ‘expected’ return. If you want to keep the higher ‘expected’ consumption path, you will have to learn to live with a less smooth consumption path during retirement.

Note that the annual amortization strategy is dynamic, conditioned on the size of the portfolio now and the expected return of the portfolio going forward, given its current asset allocation.

Conversely, the SWR strategy is static, conditioned only on the size of the portfolio at retirement and the expected return of the asset allocation at retirement. The SWR strategy does not adjust this year’s withdrawal/spending for today’s actual portfolio size or the portfolio’s current expected return. Instead this year’s withdrawal/spending is totally determined in the SWR method by
(a) what we expected the size of the portfolio to be for this year in the first year of retirement, given the expected return of the portfolio and a fixed real withdrawal amount each year. (However, this expected portfolio size could be far removed from the actual size of the portfolio this year.)
and by
(b) what we expected the portfolio to return over the entire span of our retirement back in the first year of retirement- even though in the interim we may have readjusted our portfolio’s asset allocation and/or readjusted the expected returns of the portfolio’s asset classes.

Peace, :D
Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

SWR is a bad rule - Redux

Post by bobcat2 »

Here’s one more way of seeing how the SWR rule is a bad rule. I retire at age 64 with a million dollar portfolio. Using the SWR 4% rule I will withdraw and spend $40,000 real (in terms of today’s dollars) per year from my portfolio for the rest of my life.

The stock market has three bad years over the first three years of my retirement. At age 67 after those three years of stock market losses, moderate inflation, and $120,000 in real withdrawals I have a portfolio that has seen its real value reduced to only $700,000 but, following the SWR rule, I will continue to withdraw and spend $40,000 real per year.

My twin brother retires in that third year when both he and I are both 67 years old. In that year he also has a portfolio worth $700,000 real. The SWR 4% rule says he should withdraw and spend $28,000 real per year. Adjusting for a retirement period expected to be three years shorter than mine, the SWR 4% rule advises a slight upward adjustment. This results in the rule having him withdrawing and spending about $29,000 real per year.

So here we have two 67 year old male retirees with identical $700,000 portfolios, identical portfolio asset allocations, and identical life expectancies, and the SWR 4% rule advises the one guy to withdraw and spend $40,000 real per year for life and the other guy to withdraw and spend $29,000 real per year for life.

Question: What sense does such an inconsistent policy rule make? Any reasonable strategy would advise these two guys to withdraw and spend the same amount per year in retirement going forward since they are in the exact same situation. The SWR 4% rule fails this simple test.

Of course one could argue that in this situation the first retiree should now ignore the rule. But that’s just another way of saying the rule shouldn’t be used. The problem with the SWR rule is its inability to adapt to changing circumstances. However, during retirement it’s reasonable to expect such changing circumstances when relying on income from a portfolio containing some risky assets, therefore such a static rule simply shouldn’t be used in the first place.

This is a classic case where a “common sense” rule lacks “good sense”. :(

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
User avatar
Ariel
Posts: 1361
Joined: Sat Mar 10, 2007 7:17 am

Post by Ariel »

Fascinating and instructive example, BobK. Thanks for sharing it.

Btw, a parallel argument was used by some economists a few years ago to illustrate why privatizing social security was a bad idea. In essence, if everyone had individual accounts that could fluctuate as in your example, individual payouts could differ radically despite otherwise identical circumstances.
Do what you will, the capital is at hazard ... - Justice Samuel Putnam (1830), as quoted by John Bogle (1994)
User avatar
DRiP Guy
Posts: 2241
Joined: Tue Feb 20, 2007 3:54 pm

Post by DRiP Guy »

bobcat2, I really loved your post, because it illustrates the danger of someone taking the very valuable and insightful findings of some past studyin' and possibly misapplying it (as your characters did):
("Hey, look here; if you take all stock data to date, plug in a model portfolio allocation, then using Monte Carlo modeling it looks like you could sustain up to 4% withdrawal per year over all periods of 30 yr draw down at about a 95% conf interval)
Some have taken what could then, at it's furtherest stretch, be called a "rule of thumb" -- i.e.
'when planning, a number of about 4% is likely not too far off the mark for general planning purposes, because it largely worked in all past scenarios..."
and then rigidly extrapolated it into a personal hard and fast operational constraint, i.e.
"Come what may, I'll blindly take out 4.000% because that is what Trinity said worked in past."
Working your own plan from those numbers is a little bit like dealing with annuities -- they can give you higher numbers as an annuitant because they are dealing with impersonal averages from a multitude; but for your very own one-off personal folio, it is not just some abstract concept, it is your lifeblood! Yes, it will probably work, but you probably want something like one of the many plans that adds an adaptive component to the withdrawal strategy, based on your personal balances on an 'as you go' basis. Don't have that flexibility to tailor your spending? Then work longer/save more. Pretty simple, if not cheery.

So it bothers me that so much 'talking across people' has been done on this topic.

The information and the results are not evil.

They are not "mistaken."

They are, just like much other narrowly focused research work, widely misapplied and mostly misunderstood.
Badinvestor
Posts: 801
Joined: Thu Mar 01, 2007 1:10 am

Re: One More Time - Why SWR is a bad rule

Post by Badinvestor »

bobcat2 wrote:These simulations assume that households make no adjustments to their spending regardless of how well or how poorly their investments do.
The assumption that households don't adjust their spending downwards seems to me realistic for typical retirees with up to a few million dollars saved up and no pension or former employer health care. I personally don't feel that as a retiree I will be able to cut back very much if at all on spending. Spending is dominated by health care, taxes, and housing, none of which can really be cut back (except by selling one's home and moving in with roommates, I suppose).
rower30
Posts: 29
Joined: Mon Dec 17, 2007 11:00 am
Location: Oxford, OH

Post by rower30 »

It seems to me that WHERE you save (any where in the worlds anymore) and it's savings accumulation rate is simply leveraged against your country of residents inflation and tax rate situation. So, the 4% can't possibly be a fixed ratio since all the other variables change from country to country. The same math should work, though, to figure out what you need to save to have a set amount over whatever number of years.
User avatar
segfault
Posts: 559
Joined: Thu Jun 21, 2007 4:51 pm

Post by segfault »

LH wrote: Where humans actually have a nice record of acting on thier valuations for all to see, aka managed mutual funds, the results are not pretty versus mr markets indexes.
Interesting juxtaposition, as Benjamin Graham, who adopted the metaphor of "Mr. Market," was an advocate of buying and selling stocks based upon their valuation, not of holding slices of an entire market in an index...
User avatar
gummy
Posts: 340
Joined: Mon Mar 12, 2007 2:34 pm
Location: Burlington, Ontari-ari-ari-O
Contact:

Post by gummy »

It ain't easy (for me, at least) to find historical returns for foreign markets, BUT:

If I use MSCI EAFE annual returns, namely the 33 listed here
and stick 'em into this Monte Carlo spreadsheet
and assume a 4% withdrawal rate and 3% annual inflation
I get an 85% survival rate for 30 years.

As a comparison, if I use the corresponding 33 years of S&P500
and repeat the above ritual
I get a 98% survival rate.

Ain't that interesting? Image

'course, something like Vanguard International VTRIX handily beat the S&P over the last five years.
If'n you're interested in asking Monte to compare, using just the last five years, check out the spreadsheet here:
http://www.gummy-stuff.org/2-stocks.htm
User avatar
zalzel
Posts: 557
Joined: Sat Apr 21, 2007 7:40 pm

"i", "fv", and equation relating them?

Post by zalzel »

bobcat2 wrote:So at age 65 with a 60/40 portfolio and an expected portfolio real return of 4.0% the problem becomes:

pv = -$1,000,000
i = 2.6% (4.0% minus 1.4%)
fv = 0
t = 35
payment = ? (solution about $43,900)
Hi bobcat2.

Would you please define "i" and "fv", and provide the equation that relates these to "pv", "t" and "payment"?

Thanks.
hafis50
Posts: 332
Joined: Fri Jun 22, 2007 1:55 am

Post by hafis50 »

gummy wrote:It ain't easy (for me, at least) to find historical returns for foreign markets, BUT:

If I use MSCI EAFE annual returns, namely the 33 listed here
and stick 'em into this Monte Carlo spreadsheet
and assume a 4% withdrawal rate and 3% annual inflation
I get an 85% survival rate for 30 years.

As a comparison, if I use the corresponding 33 years of S&P500
and repeat the above ritual
I get a 98% survival rate.

Ain't that interesting? Image

'course, something like Vanguard International VTRIX handily beat the S&P over the last five years.
If'n you're interested in asking Monte to compare, using just the last five years, check out the spreadsheet here:
http://www.gummy-stuff.org/2-stocks.htm
But didn't you calculate the returns and SWRs for US investors in foreign markets?

I believe the original question referred to the SWRs for foreign investors who invested in their home countries.

Another calculation would compare the SWRs of global portfolios (e.g., the MSCI World Index; it may be difficult to find older data) in their different home currencies.
mikenz
Posts: 780
Joined: Sat Mar 10, 2007 7:33 pm

Post by mikenz »

Hi bobcat2.

Would you please define "i" and "fv", and provide the equation that relates these to "pv", "t" and "payment"?
in Excel there is a function PMT( rate, nper, pv, fv ) where
rate: interest rate
nper: number of periods/payments
pv: present (starting) value (usually input as a negative number)
fv: future (ending) value

In this case rate is 2.6%, nper is 35, pv is -1,000,000 and fv is 0 (end up with nothing).

So the formula =PMT( 2.6%, 35, -1000000, 0 ) gives you $43,862

There are also functions =PV(..) =RATE(), =NPER(..), so for instance if you want to calculate how long 1,000,000 would last at 2.6% withdrawing 43,862, you can use =NPER( 2.6%, 43862, -1000000, 0 ) and get 35.
User avatar
zalzel
Posts: 557
Joined: Sat Apr 21, 2007 7:40 pm

PMT function

Post by zalzel »

mikenz wrote: in Excel there is a function PMT( rate, nper, pv, fv ) where
rate: interest rate
nper: number of periods/payments
pv: present (starting) value (usually input as a negative number)
fv: future (ending) value

In this case rate is 2.6%, nper is 35, pv is -1,000,000 and fv is 0 (end up with nothing).

So the formula =PMT( 2.6%, 35, -1000000, 0 ) gives you $43,862
Thanks mikenz.

So... i of 2.6% in the illustration, is the Expected Return of 4%, minus an arbitrarily chosen 1.4% safety margin?

Would you (someone) please state the "PMT" function/equation for those who do not use Excel?

Thanks.
User avatar
zalzel
Posts: 557
Joined: Sat Apr 21, 2007 7:40 pm

Gummy's illustration

Post by zalzel »

gummy wrote: If I use MSCI EAFE annual returns, namely the 33 listed here
and stick 'em into this Monte Carlo spreadsheet
and assume a 4% withdrawal rate and 3% annual inflation
I get an 85% survival rate for 30 years.

As a comparison, if I use the corresponding 33 years of S&P500
and repeat the above ritual
I get a 98% survival rate.

Ain't that interesting? Image
Hi Gummy.

Are you not including a Bond allocation in these portfolios? If you are, what is the allocation and what return series are you using? Would you mind re-running your simulation for a portfolio containing equal allocations to U.S. and EAFE equities?

Thanks.
User avatar
gummy
Posts: 340
Joined: Mon Mar 12, 2007 2:34 pm
Location: Burlington, Ontari-ari-ari-O
Contact:

Post by gummy »

But didn't you calculate the returns and SWRs for US investors in foreign markets?
Yes.
Would you mind re-running your simulation for a portfolio containing equal allocations to U.S. and EAFE equities?
There's a spreadsheet described here which will let you do that:
http://www.gummy-stuff.org/2-stocks.htm
User avatar
artthomp
Posts: 327
Joined: Tue Feb 20, 2007 3:44 pm
Location: Ballwin Missouri

IRS RMDS as annuity

Post by artthomp »

The IRS Required Minimum Distribution can be used to illustrate a Safe Withdrawal Rate if only the correct percentage is withdrawn each year. If the principal declines and only the Required Minimum Distribution is taken, the money will never basically run out. In very severe conditions the amount of money available might become much less than expected however (Japan case). This would also be the case if only 3 or 4 percentage were always taken out each year based upon the current principal.

Required Minimum Distributions viewed as an annuity

IRS rules require that a certain amount of a traditional IRA be withdrawn each year after the IRA owner reaches the age of 70 ½ years.

The distribution rules were made more generous recently and it looks to me like it works better than most annuities. If only the IRS Required Minimum Distribution (RMDS) percentage is taken in bull or bear market years, the amount of money withdrawn will be increased some years and reduced some years but will theoretically never run out. Of course a severe long term decline like the Japanese scenario might make the amount of money insufficient to support required retirement expenses. Portfolio allocations for retirees dependent upon investments for a significant portion of their living expenses should be conservative anyway (higher percentage of fixed income reducing the equity risk).

Assuming the portfolio grows in value 6%/year (The default value in the Vanguard planner), the portfolio grows 19.5% in value after the distributions for the first 13 years. If the recipient is lucky enough to live to 100 years, the portfolio would have lost 37% of its initial value.

Since the distribution is based upon the recipient’s life expectancy, it increases each year as the life expectancy decreases. For example the distribution for the first year is based upon a life expectancy of 27.4 years. If one makes it to 100 years old the distribution is based upon a life expectancy of 6.3 years. This means the distribution income doubles after 13 years and peaks at 288% of the initial distribution value when one reaches 96 years of age. It is still 270% of the initial distribution value if one reaches 100 years of age.

Distribution from the IRA only means that taxes are due. After taxes are paid, if some portion of the distribution is not required for living expenses, it can be reinvested outside the IRA.

A final important consideration is that a surviving spouse can roll an IRA account from a deceased spouse directly into his or her own IRA account. Of course any amount remaining after both spouses are dead goes into the estate but can be rolled over into the descendants IRAs. Many annuities only guarantee payment while the recipient is living.
Art
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

zalzel

Post by bobcat2 »

These are the financial workhorse Time Value of Money (TVM) calculations. As other posters have noted these problems can be solved using financial functions in spreadsheets. They can also be solved on inexpensive financial calculators. The important point, however, is that the calculations need to be done every year to keep your future spending path consistent with the current size and current expected future return of your portfolio. For that matter you could run the Monte Carlo simulations every year, but personally I don't think that extra effort is worthwhile.

What does not make sense is doing either calculation once (at the beginning of retirement) and then never redoing the calculations. This static non adaptive approach is a recipe for poor to bad results if actually followed. Unfortunately, doing it once and not making any adjustments over time is the conventional advice. This is particularly bad advice if your portfolio consists of 50% or more risky assets.

The goal should be to keep your living standard (consumption) as high and relatively smooth as possible during retirement, not simply running out of money (zero consumption). That goal sets the bar far too low. Withdrawing and spending $10,000 per year out of a million dollar portfolio would accomplish that. So what? :roll:

Bob K

PS - Without a safety margin you are basically looking at the equivalent of a Monte Carlo simulation with a 50% success rate.

PMT = periodic payment
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Ariel and DRiP Guy

Post by bobcat2 »

Ariel,
Yes, there are some similarities between this and the SS arguments. I hadn't thought of that.

DRiP Guy,
I would add to your advice of working longer and saving more to also at a minimum think about the timing of beginning SS benefits, considering a reverse mortgage, and deciding how much of the retirement portfolio to annuitize.

Unfortunately one often used strategy is the realization that the plan doesn't work with the current 50% equity, but with the higher expected portfolio return from 80% equity the plan works. Therefore I will solve the problem by moving to 80% equity. If you implement this 'solution' you need to have a solid 'Plan B' in place. :D

Bob K
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
ajbibi
Posts: 203
Joined: Mon May 07, 2007 10:27 pm

Post by ajbibi »

bobcat,

I appreciate the virtues of consumption smoothing, but I have a comment:

If you don't know what the retiree's utility function is, all you're doing with the safety margin is building in a guess about risk aversion/preferences into the rate of return.

After all, if I choose a high enough safety margin, I will still end up with a large unspent surplus at death.

The goal of consumption smoothing is entirely different from the problem of what rate of return to model.

As far as I can tell, the SWR is a crude rule of thumb for not running out of money.

The consumption smoothing says: Given a preferred time profile of consumption, and given a good guess estimate of returns here's what you should be spending now. Redoing it every year is just a more sophisticated version of the variable/dynamic withdrawal rates. But at the heart of it is still an assumption about returns. But now you're adding assumptions about people's utility preferences with regard to spending changes. Granted, the Kotlikoff (or whomever's) model incorporates more reasonable assumptions about utility, the SWR model works well under the conditions that 1) you're interested in loss minimization and 2) you want a fixed, inflation adjusted draw with large probability of ex post surplus.

Granted, this is not the most sophisticated analysis, but I don't see that -- as a rule of thumb -- it's much worse than someone using (for example) the market annuity rates from Vanguard as a benchmark for how much consumption their portfolio will support.

BTW, I've played with Esplanner now, and I'm impressed by it. Nonetheless, I don't see how the average person -- without training in econ or finance -- will understand what it means to change around various inputs and assumptions. In my own case, I could make the recommendations jump around a lot, and it took quite a bit of work for me to figure why some of the changes produced different recommendations. And only after many tries did I get to a range that I understood and was comfortable with.
User avatar
VictoriaF
Posts: 20122
Joined: Tue Feb 27, 2007 6:27 am
Location: Black Swan Lake

Post by VictoriaF »

ajbibi wrote:BTW, I've played with Esplanner now, and I'm impressed by it. Nonetheless, I don't see how the average person -- without training in econ or finance -- will understand what it means to change around various inputs and assumptions. In my own case, I could make the recommendations jump around a lot, and it took quite a bit of work for me to figure why some of the changes produced different recommendations. And only after many tries did I get to a range that I understood and was comfortable with.
Thanks to Bob (bobcat2), I was thinking of getting ESPlanner to experiment with. Now, I wonder if my lack of economics and finance background would make it less instructive. I usually like to understand what produces different results, and in this case, this may be even more important.
Inventor of the Bogleheads Secret Handshake | Winner of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)
User avatar
bobcat2
Posts: 6074
Joined: Tue Feb 20, 2007 2:27 pm
Location: just barely Outside the Beltway

Post by bobcat2 »

Hi ajbibi,

The primary problem with the SWR is that it doesn't adapt to changing circumstances. See my above example of the two twins. In that example the twins at age 67 are in exactly the same financial situation. Yet the SWR gives wildly different solutions. This is not reasonable.

Just because you have only a 5% probability of outliving your portfolio assets with a particular strategy at the time of your retirement doesn't mean that probability stays constant if you don't adjust the strategy for changed circumstances. It's possible that 10 years into retirement that strategy might have a 25% prob of your outliving your portfolio resources or, on the other hand, a 0.5% probability of outliving your resources. Presumably what you want to keep constant is the 5% probability, which means adjusting the strategy over time. SWR strategies do not do that!

The reasonable goal in retirement is to attempt to keep consumption relatively high and relatively smooth. It's true that individual utility functions will determine the trade-off between higher expected consumption and lower but smoother expected consumption. This is a reasonable goal and trade-off that you have to make regardless of whether or not you use ESPlanner to help make these decisions.

Bob K

PS
I certainly agree that ESPlanner is not a tool that will be useful for most people to use directly. It does require a good understanding of both financial and economic concepts. On the other hand people reading this deep into a Boglehead thread on SWR's are for the most part far removed from the typical American analyzing retirement strategies. :D
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.
ajbibi
Posts: 203
Joined: Mon May 07, 2007 10:27 pm

Post by ajbibi »

bobcat,

The point I was making was that dynamically adjusting to different circumstances is something that you can do without smoothing. For example, you can recalculate your SWR each year so that you get 4% of 1M or 4% of 700k depending on how well your fund has done. That allows for annual adjustments. But this dynamic adjustment is *different* from smoothing.

Using ESPlanner to recalculate your spending each year combines the functions of smoothing and dynamic adjustment. Granted, most people want smoother spending, but knowing what kind of smoothing to apply is different from dynamic adjustments.

To take just one example, do you want a fixed smoothing profile or a rising standard of living? Do you use a fixed adjustment for inflation or do you use a personal CPI? [e.g. it doesn't matter if US CPI is 3% if the CPI for your personal bundle is 5%] Do you suffer from money illusion? Some people might be happy with a 2% real rise when inflation is 3% but a zero real increase when inflation is 7%. Are you more sensitive to nominal vs real declines [Some dynamic withdrawal strategies have you withdrawing a fixed sum when markets decline and half of %gains when markets rise]? Being flexible in consumption will inherently improve the chance of your portfolio surviving and allow a higher initial draw.

If you're willing to adjust your consumption dramatically every year, smoothing may not be necessary.

[More technically, most people -- including economists -- don't know the shape of their utility function and have only a vague idea of how to factor that into their consumption profiles.]

I was thus making a minor but important point: You can want smoothing and still get wildly inaccurate estimates of return or you can have return risk and not much care about smoothing. The best smoothing algorithms don't protect you from "underconsuming" if you assume too low a rate of return. Conversely, assuming too high a rate of return will have you consuming too much in the beginning, in rough analogy to picking too high an initial draw.
Post Reply