Smart guys question 4% SWR strategy

[aenlighten]

the biggest flaw with this is trading success over 30 years for failure over 31. Not what I would call lowering risk, but trading life expectancy risk for portfolio risk. Now they do suggest that reducing equity exposure later in retirement helps to address this, and taking investor investment returns and spending preferences into account are important, both of which are likely true. I think most people do take these into account, at least over the long term, and the 4% rule is really just a guide. The more I look at the 4% rule, though, the more robust it appears.

[Nords]

It's a good thing these guys have their chops. A cynical reader would be inclined to call it "deadline journalism research".

This is a great survey SWR's best literature, these guys clearly had fun with it, and I'm a sucker for their Boomer parable. (Nice reference bibliography for future SWR researchers, too.) However they made gross assumptions to simplify their math. And if they're not gonna propose a solution then why whine about the status quo for 24 pages?

First, page 8 assumes that investment returns are log-normal. In the real world, they're not. The real bell curve is a lot flatter and has black swans much fatter tails.

Second, they assume that equity returns are both independent and serially uncorrelated. In the real world, again they're not. Next year's returns are usually a persistence of the following year's returns. ("The stock market rises two-thirds of the time.") These assumptions are among the reasons why Monte-Carlo simulations are more conservative than historical-- MC randomly scatters good & bad years instead of letting the good years run.

Next, the authors sell nonexistent products like perfect T-bills, constant inflation, and capless COLA'd annuities. They cheer over a 4.46% payout on page 8 and they're accounting for inflation by using "real dollars". (They're assuming that inflation stays at some constant for 30 years.) That's another assumption that greatly simplifies the math but has no semblance of reality. If you can find an affordable COLA'd annuity, the insurance company hedges their risks by capping the COLA. Admittedly that 8% or 10% cap rarely kicks in, but then these authors are already complaining that a 4% SWR isn't optimal. Over a 30-year annuity there would be one or two times when inflation might exceed the cap, but no survivable insurance company would sell a capless COLA'd annuity.

These guys also ignore investor psychology. Investor's won't blindly waste the excess ("Woo-hoo, the market's up! Spend it all, dammit!!") but instead will almost certainly carry some over to next year's spending. Next year they won't withdraw as much, improving the portfolio's survivability. In down years there will also be some deferral of spending.

It'd be interesting to see an SWR model that also accounted for reduced discretionary spending starting around age 75-80, as well as rising health insurance expenses (like Fidelity's $200K of healthcare costs near the end of life).

I guess we shouldn't be surprised that Sharpe prefers to annuitize spending. That's how FinancialEngines.com is set up, too.

Finally I agree with the poster who feels the authors crapped over all the research without suggesting a better method. So we're "wasting" money by achieving a minimum lifestyle and gambling the profits on a better one. Should we have retired earlier? Lived lower below our means? Annuitized the majority of our portfolios to remove all uncertainty? With so many approximated variables and so many unquantified investor-psychology characteristics, isn't it likely that the average ER has already put enough slack in the system to be able to gamble an occasional buck?

Luckily those guys have no plans to retire in the first place. I wonder if they're going to use this paper as the basis of a new insurance company selling fully-COLA'd 30-year annuities for ERs who don't want to "waste" any of their money...

[Texas Proud]

I read a lot of the article.... and as some of the people said... not quite right..

There is a BIG difference in making a decision for a perfect $125 ticket that you will use NOW and trying to predict how much that ticket will cost in 30 years and investing NOW to have the money THEN...

And as we know.... what if the concert is not for 40 years

The reason we invest to have more and 'cost' us more money or have more return than we 'need' is because of all the unknowns.... if they were all known, it would be easy...

[cute fuzzy bunny]

Uncertainty is the spice of life.

[HFWR]

Originally Posted by cute fuzzy bunny

Uncertainty is the spice of life.

I'm not so sure about that...

[Nords]

Originally Posted by cute fuzzy bunny

Uncertainty is the spice of life.

Wouldn't that lead to indigestion, gastric ulcers, and bazooka puking esophageal reflux?

These days I feel that we're somewhere between "Thai curry" and "Texas chili"...

[unclemick]

Originally Posted by cute fuzzy bunny

Uncertainty is the spice of life.

A little boredoom north of Kansas City for a couple years has been ducky.

Now I 'could' cross the river to Kansas in search of the yellow brick road - or take a trip to New Orleans the end of the month for the Jazz and Heritage Fest.

The stupid tornado siren in my neighborhood is only a few blocks away - loud when it goes off.

heh heh heh - the financial part of uncertainity seems easier somehow.

[cute fuzzy bunny]

Eh, if you aint worried about anything, you aint livin!

[rogersteciak]

Originally Posted by ESRBob

Yes -- in the first round of studies we just used US data, but in the second round I had access to the Dimson Marsh Staunton series of historical international equities data. The result was that the portfolios with the international data fared a little worse, but not materially. It could be because of your point that the 4 horsemen have shown up in a lot of places during the 20th century, trashing markets in their wake, and that perhaps they haven't shown up in as much force or frequency in the US keeping our market performance on the upper end of international norms. As you say, it could be our turn again, at which point we'll be glad to be globally diversified!

This is good news. The fact that the Work Less, Live More approach to early semi-retirement works with global stock market data (which has a lower performance overall than the U.S.-only data for a variety of reasons) gives me greater confidence in the approach.

I subscribe to the Yogi Berra school of predicting the future (i.e., "the future ain't what it used to be"), so there is a margin of safety with the Work Less, Live More approach should the U.S. markets happen to revert to the global mean at some point in our lifetimes.

[cjking]

I think someone made the point that you can't get 2% on inflation-linked bonds at the moment. I did the calculation, and even if they are paying 0%, that only lowers the risk-free withdrawal rate from 4.46% to 3.33%. In other words, most of the income comes from running down capital anyway.

This reminds me of a section in "Valuing Wall Street" where they are arguing that the benefits of shares have been oversold, and show that the difference in investment returns between a high-risk asset and a safe one makes very little difference to the income a retiree gets. For example, for a hypothetical annuity based on equity returns they calculate that someone with a 25 year life-expectancy would receive a 15.1% higher income than if their annuity were based on index-linked bonds.

[cute fuzzy bunny]

Its only "risk free" if your personal inflation rate matches or is lower than the CPI.

Plus you still have to die on schedule.

[cjking]

I've been working out how to take on board the lessons of this paper. I think it's easy. You start by following the basic "4% strategy". You stay 100% invested in "the market portfolio", while taking income of 4% of your initial capital, adjusted for inflation each year. If/when your capital is sufficient to buy the same income from safe assets for the rest of your retirement, you switch 100% of your capital into those safe assets. There are no unused surpluses. Utility is maximised, in that you have the same real income over the whole retirement period, provided you don't run out of money.

I've deliberately left terms like "inflation" and "the rest of your retirement" and "safe assets" not to-precisely defined. They can mean what you want them to.

For me "safe assets" would have to mean an increasing income from an annuity, and the "inflation" I would like to see my income increase by each year would be the rate of growth of individual earnings. The happiness money can buy depends on your income relative to the rest of the population. It's not good enough to keep up with even your own personal inflation rate, to maintain constant happiness you need income that increases in real terms.

I put "4% strategy" in quotes because for various reasons 4% is not always the right number. Wanting an income that grows in real terms, as I do, would depress it. For me, stock-market valuation on the day you retire also affects the number. I calculate my expected real return for the stock-market at the peak of the year 2000 boom as being about 2%, compared to 5.7% now.

[ERD50]

Originally Posted by cute fuzzy bunny

Its only "risk free" if your personal inflation rate matches or is lower than the CPI.

It's certainly true that there is risk that your personal inflation rate can vary from the CPI.

But that is true no matter what the source of income is, so that 'risk' does not seem to play into decisions about the investment side.

-ERD50

[samclem]

I think there might be some holes:

Originally Posted by cjking

I've been working out how to take on board the lessons of this paper. I think it's easy. You start by following the basic "4% strategy". You stay 100% invested in "the market portfolio", while taking income of 4% of your initial capital, adjusted for inflation each year. If/when your capital is sufficient to buy the same income from safe assets for the rest of your retirement, you switch 100% of your capital into those safe assets.

Implications:
1) If the value of your portfolio never rises to the point where you can buy the required income flow from these "safe" assets, then what?

2) "Safe Assets": Without defining this, it's not clear what you are talking about. For example, if you wanted to convert to this income stream today, what assets would you buy?

Originally Posted by cjking

For me "safe assets" would have to mean an increasing income from an annuity, and the "inflation" I would like to see my income increase by each year would be the rate of growth of individual earnings.

So you want an annuity that increases the monthly payout as long as you live at the same rate as U.S. individual earnings. Are such products available, and at what cost?

[cjking]

Originally Posted by samclem

I think there might be some holes:



Implications:
1) If the value of your portfolio never rises to the point where you can buy the required income flow from these "safe" assets, then what?

Then you are stuck in the standard 4% strategy. I'm not convinced there's anything you can do. For example, whenever I've looked at using options to limit volatility in the past, I've concluded they're not worth bothering with. If anyone has ideas on how to dynamically implement a strategy that achieves the effect of their sell-calls-at-one-strike-and-buy-them-at-a-lower-strike strategy, I'd be interested to hear. (Their actual strategy is impossible because you can't actually trade 30 year calls.)

Setting aside for the time being the possibility of doing something clever with options, I think if you don't have enough capital to switch to safe assets, then you don't actually have a problem with surpluses. No matter how much your investments have risen, what you have is not a surplus but a potentially inadequate buffer against the effects of a downturn. You shouldn't use that money for increased spending from your current target, as it has a more important function of ensuring your spending doesn't fall below you current target at some future date.

2) "Safe Assets": Without defining this, it's not clear what you are talking about. For example, if you wanted to convert to this income stream today, what assets would you buy?

Either inflation-linked government bonds or a inflation-linked annuity.

So you want an annuity that increases the monthly payout as long as you live at the same rate as U.S. individual earnings. Are such products available, and at what cost?

No they aren't. I haven't entirely thought this through, but the way I think I'd achieve this is to define a safe income as an inflation-linked annuity that pays more than my target spending level, with the excess being reinvested in "safe assets."

[OAG]

IMHO this is gets overcomplicated (even tho those Excel Spreadsheets can become a lot of fun to play with, out 30 years or so). IMO 4% is a very good number to ensure the money lasts (but is can become too restrictive sometimes). If you have $2MM on Jan 1 in money assets (forget the house) you could safely spend $80K that year. If you still have $2MM in money assets on the next Jan 1 you can do the same the following year. If not, 4% of what it is would be safe. If it ever gets to the point the 4% will not get you through the next 12 months maybe then you have to do something else. Additionally, while inflation is important, it can become the "ghost in the closet" and be feared too much. Having said that, the adjustment to the current economy, which is ongoing, may change a lot of things in the near to long range future of many people.

[cjking]

Originally Posted by R Wood

IMHO this is gets overcomplicated. IMO 4% is a very good number to ensure the money lasts. So, if you have $2MM on Jan 1 in money assets (forget the house) you could safely spend $80 that year. If you still have $2MM in money assets on the next Jan 1 you can do the same. If not 4% of what it is would be safe. If it ever gets to the point the 4% will not get you through the next 12 months maybe then you have to do something else. The prudent would probably not spend 4% each year and what is left over could be saved or spent the next year (in addition to the "normal" 4%).

What you are describing is a safer strategy than the standard 4% strategy they criticise in their paper. The 4% strategy they criticise is the one where you take 4% of the capital you had at the start of your retirement, regardless of what happens to your investments thereafter. Unlike your version, that strategy can run out of money.

Your version still has some the problems though. Income is variable, so you are not maximising utility. (Each extra dollar you spend is assumed to buy things you value less than the one before, so if you spend $110 one year and $100 the next, $5 you spent in the earlier year is sub-optimal because with hindsight it would have bought something you wanted more in the second year.)

Also, if your investments do well and you don't buy an annuity at the right time, you will end up with a surplus that you can only spend on things that are less valuable to you than guaranteeing you don't end your days eating cat-food. Actually I think this is just the same problem I mentioned in the previous paragraph. The real problem with never buying an annuity is either running out of money or leaving money unspent, in this case only the second applies.

[samclem]

I think you'll find, once you boil your idea down, that it will consist of gradually shifting to less risky assets if your portfolio is successful enough to allow these assets to support your lifestyle. I think most people would would agree with that approach.

Of course, keeping a well diversified portfolio of risky but uncorrelated assets allows an investor to reduce overall portfolio volatility while maximizing return. I would feel much safer with such a basket than with a multi-decade annuity backed by an insurance company.

Of course, this does not address the issue of dying with money left over. That's a "problem" I'm happy to have. It will give me a chance to do some good for others, which I think is a better option than helping Prudential build another skyscraper. And those, I think, are the two options.