Why rebalance in retirement?

[pb4uski]

Not sure that it matters much.... first is $1m in a 60/40 with $40k inflation adjusted withdrawals with annual rebalancing, second is same scenario but with no rebalancing.

Similarly, FIRECalc suggests similar levels of success across a broad range of AA.

So while taken together it doesn't seem to matter much, I do support periodic rebalancing.

 

[MichaelB]

Not to provoke arguments, but IMO the idea that volatility is risk is waay oversold.

Of course you’re provoking an argument. We’ve disagreed in the past on this, nothing has changed, so no need to repeat the same arguments. I’ll move on.

 

[MichaelB]

Fine ( though I agree with OldShooter that 'risk' and 'volatility' are not the same thing - if you lower your 'risk/volatility' to near 0 with all cash, your portfolio won't survive for long, and that's the real 'risk' we are concerned with), but that is separate from a statement that rebalancing improves the portfolio survival rate through lower volatility (which is what I questioned).

-ERD50

I didn’t say risk and volatility are the same thing and I certainly never suggested reducing volatility to zero. In fact, I haven’t suggested minimizing volatility. That’s a red herring.

Just to summarize, once again, in different words. I think setting an asset allocation target establishes an expected or tolerable level of risk. Allowing an equity allocation to remain above the target level increases portfolio risk and increases volatility. Rebalancing a portfolio reduces risk by returning the allocation to target levels. Increased volatility contributes negatively to portfolio survival. These factors are probably less significant at the core and have greater impact at the extremes.

 

[steelyman]

I agree with MichaelB’s viewpoints. I think it’s an individual choice regarding rebalancing approaches. I do it through annual withdrawals. I’m not interested in shooting the lights out (as my reports in the YTD Performance thread will attest) but rather to keep a smooth sailing, steady as she goes behavior for my retirement portfolio.

 

[RobLJ]

Many here have a very low withdrawal rate (or with pension/SS don't even depend on the portfolio) and wish to maximize their portfolio.

However, as the 2000 crash indicated, "volatility is risk" (in the form of a severe extended market downterm before recovery), at least for those making significant withdrawals before pensions or SS withdrawing, and who are withdrawing at a significant withdrawal rate. Not for those who are withdrawing nothing or less than 1.5/2% from their portfolios either because of low spending/thrift or other sources of $.
However, if you include the 2007 crash, it could have taken as long as 14 years to recover., even not considering withdrawals. I was unaffected largely because I was in accumulation mode and dumping a large % into stocks, until 2005 when I started bringing down my allocation from 90% stocks, although I did bump stock investments then to rebalance back up. For those at WR and at WRs of 2.5% or even 3% or less, this period would be a lot less significant than someone taking 4-5% WR and 80-100% stocks. A further factor is the psychological--will someone at 90% stocks and withdrawing 4% or 5% before SS in, say 8 years, be able to stick to that allocation until recovery at, say, 2012 or 2013? Many here would I guess, based on testimonials. I doubt I would, so a lower stock allocation that might harm my "ultimate" portfolio when I croak or even reduce spending 2-5% doesn't seem the correct goal, at least for me.

I am willing to let my stock allocation creep up as I get closer to SS FWA, although I'll periodically scrape stock gains over 55% allocation for sure, either during the year or during the Jan withdrawal.
And yes, one could work longer, save more and use a lower WR rate to "maximize" the ending portfolio, but this is a ER forum.

I'm attempting to maximize withdrawals and reduce SORR risk in the period over the next 7 years before my and DW's FRA/SS and am not trying to maximize the amount of $ left over when I/DW croak. This is a different goal than many, but one--given SORR risk--both rebalancing and a lower stock allocation lessen the short-term risk, no doubt at the expense of maximizing spending later. particularly "normal" stock return environments. Another fix is a variable withdrawal rate, as others have posted, particularly if you have a large "optional spending" buffer.
In my case, I am holding cash sufficient for the next 3.5 years until my SS FWA and will use bonds/rebalancing for the additional 4 years until DW's FWA. On my FWA and certainly after DW's the risk of SORR considerably diminishes, so at that point I will gradually increase the stock allocation. Could I squeeze out another 1-2% in bond income from cash--yes, but is it worth the trouble with treasuries at 1%?

Until then I usually can rebalance with January withdrawals, although I did shove more into stocks in March/April of 2020 (it seemed a good moment and one when stocks were cheaply valued compared to previous years, and I was prepared to use more cash if it continued to decline).

YMMV, particularly if your goal is to maximize a portfolio for heirs or charity, which are worthy portfolio goals. Unless we run into another 2000/2007 scenario (which I suspect is close to a worse case for the S&P), I doubt it will make much difference, as several have posted. But in one of the bad scenarios--it could make a difference and it is those bad lines at the bottom of the FireCalc chart I worry about; otherwise I would be withdrawing 7% a year.

 

[chassis]

In all things business and investing, I like measurable KPIs. I know many people on this site were/are professional managers, executives and business owners. How many of you don't measure anything in your business, quantitatively? I'll bet not many of you.

So let's be precise about risk. Is it:

- standard deviation of return
- conditional value at risk
- maximum drawdown

Or something else? Let's be specific. Otherwise it's chasing after the wind when risk is not defined. All three of the above are good risk KPIs in my book.

 

[RobLJ]

Given my post, to me it is a combination of max drawdown, extended over a long period, as my post suggests.

I agree there are other primary risks, including long-term inflation, for example. This VERY MUCH is determined by one's withdrawal rate, portfolio goals, longevity, etc, etc, so I do not suggest that my #1 risk is or should be your #1 risk. SORR risk probably only applies to higher withdrawal rates (>3.5%) and longer times to pension/SS, and only in the bottom 10-20% of the FireCalc curves, so the maximize portfolio posts will in all likelihood turn out to be right (which doesn't bother me).

 

[Markola]

Another risk is Retiree Rigidity, if you will:

“…[T]he key to choosing your own (withdrawal) rate has less to do with the numbers than with your personal flexibility. If as needed you can readily adjust your living expenses, find work to supplement your passive income and/or are willing and able to comfortably relocate to less expensive places, you will have a far more secure retirement no matter what rate you choose. Happier too I’d guess.

If you are locked into certain income needs, unwilling or unable to ever work again and your roots go too deep to ever seek out greener pastures, you’ll need to be much more careful. Personally, I’d work on adjusting those attitudes. But that’s just me.”

JL Collins

https://jlcollinsnh.com/2012/12/07/stocks-part-xiii-withdrawal-rates-how-much-can-i-spend-anyway/

 

[SnowballCamper]

I started to write this yesterday, but didn't post. We're being awfully vague about risk.

A risk is an adverse future event with it's associated probability.

So, of course, volatility is not risk.

The risk is being forced to sell depreciated securities for essential spending due to a market downturn. The adverse event is the forced selling, and the probability is of the market downturn happening at the same time you have to spend invested $$ to live.

Inflation alone is not risk. The risk is that your essential spending amount exceeds your investment returns (over some finite time period) with the probability determined by both future inflation and the investment returns.

Of course in most cases the probability part of risk is unknown, and therefore a subjective quantity we get to make up.

I just think it would be helpful if, when someone wants to post about a risk, they could be a little bit clearer and identify the adverse event they have in mind along with how probable they think it is.

Ok, my rant is over....sorry to bother you (not really ;-)

 

[chassis]

I started to write this yesterday, but didn't post. We're being awfully vague about risk.

A risk is an adverse future event with it's associated probability.

So, of course, volatility is not risk.

The risk is being forced to sell depreciated securities for essential spending due to a market downturn. The adverse event is the forced selling, and the probability is of the market downturn happening at the same time you have to spend invested $$ to live.

Inflation alone is not risk. The risk is that your essential spending amount exceeds your investment returns (over some finite time period) with the probability determined by both future inflation and the investment returns.

Of course in most cases the probability part of risk is unknown, and therefore a subjective quantity we get to make up.

I just think it would be helpful if, when someone wants to post about a risk, they could be a little bit clearer and identify the adverse event they have in mind along with how probable they think it is.

Ok, my rant is over....sorry to bother you (not really ;-)

@SnowballCamper can you propose a quantified measurement for your view of risk?

You talked about selling in a market downturn. The title of this thread includes the word "rebalance". In your view how are rebalancing and selling in a downturn related?

 

[SnowballCamper]

Sure, the risk matrix is the common way of quantifying risk. https://en.wikipedia.org/wiki/Risk_matrix. And no, it doesn't result in a number, because it relies on the probability of a future event, which is not precisely known. Assigning a number would give a false sense of precision.


In a stock market downturn the prudent thing to do is to rebalance (sell stable or appreciated assets and buy more stock at the now lower price), not to sell stocks. On the one hand, this seems obvious, but on the other hand, the lower stock prices are precisely the result of more willingness to sell stocks, even at the lower price which is selling in a downturn.

 

[OldShooter]

... a quantified measurement for your view of risk? ...

Just for reference, the genesis of the idea that volatility is risk is AFIK Harry Markowitz' 1952 paper, "Portfolio Selection," which was the first inkle of modern portfolio theory (MPT eventually earned him a Nobel). He had failed to find a quantitative measurement for risk and he needed one to do the mathematical analysis he wanted to do. So rather than explaining, he simply postulates: "We next consider the rule that the investor does (or should) consider expected return a desirable thing and variance of return an undesirable thing."

Starting with variance is kind of an odd pathway to "risk" equivalence because variances can be both positive and negative. But for him, an accurate prediction of return was the goal, so variance in either direction was the undesirable thing. (William Sharpe later patched over this up/down thing with his "Sharpe Ratio.")

To me risk is like Justice Potter Stewart's approach to pornography: "I know it when I see it." Risk is Enron, Theranos, Montgomery Ward, GE, Worldcom, Compaq, Kodak, etc. (Many times hindsight tells me that I have looked at risk without seeing it, but that is another thread.)

But @chassis, your implicit point, maybe a challenge, is that measuring risk quantitatively is difficult, maybe impossible. I agree but will be pleased if @SnowballCamper has one.

(@SnowballCamper identified the one place, though, where variance creates risk -- forced selling of volatile assets. SORR, in other words.)

 

[chassis]

@OldShooter risk can be easily quantified. It depends on the definition of risk. It's a very broad topic that can and should be partly quantified. On the contrary, there is no implicit question, unless you manufactured the question yourself. Your posting style seems confrontational, seems to seek conflict and seems to propose words on behalf of others. It's OK, it doesn't bother me, my skin is pretty thick. But definitely something I have noticed and I think others have commented about your posting style as well.

The aspects of risk that I pay attention to in the investing world are in the next sentence.

I like standard deviation of return, max drawdown, downside capture ratio, value at risk and intra-portfolio correlation.

Examples you gave fall into fraud and mismanagement in my view. What about meteors hitting the earth? Serious question. What about airliners flying into tall buildings in a global financial center? What about a global panic lasting nearly 2 years?

Too many scenarios to identify, so let's boil it down to some quantified KPIs.

 

[SnowballCamper]

I agree that there are too many high level scenarios to identify. The one scenario that matters to the individual investor is being forced to sell a depreciated asset in order to fund essential spending, or just running out assets in general. I'm not sure how the various KPIs that chassis mentions relate to those two events in a direct manner...even taken as a whole. They seem to me to be in line with "volatility is risk" thinking.

The classical statistical way to quantify a financial risk would be with an expected value (the average of the product of each each $$ outcome and its associated probability) but we quickly realize that single number is not very descriptive of the possible outcomes. Then a more robust approach would be to use the whole distribution of outcomes. But it turns out that we don't really know the distribution in many cases, or make shortcut assumptions so that the math is easier. This making shortcut assumptions is necessary for the individual investor, and indeed it is what everyone who uses FIREcalc does in interpreting the spaghetti graph of historical returns. We assume that the historical returns are an indicator of future results.

So how would snowball measure risk with a number? If forced to use a number, I would abandon the combination of event and probability, as the above paragraph describes the complicating math, and esoteric assumptions. My initial thought of how to measure personal financial risk is in units of $$:

Current Assets (to include vested pension & Social Security streams) - Market realized losses in the last downturn (maybe times 2 if under 50y/o) - Current annual average spending*(100-age)

The resulting number (if negative) is how much gain you need from investments (or work) in order to maintain current spending levels through one market downturn (or two) like the last one, and live to be 100. If the number is positive, then you can give all your market gains to charity.

This is a very crude measure, that ignores a lot, but I think it captures the essential idea of quantifying an individual's risk of running out of money before death. It does so in a most useful way, by telling you how much money you will need to come up with to maintain your current spending until age 100.

It doesn't rely on assumptions about market returns. But it does take into account the individual's portfolio performance in the last downturn. People tend to be slow to change behavior...

It does require the user to make assumptions in order to determine current assets, and about future spending, surely not too much of a cognitive load.

So this is just what I came up with in the last six hours since Oldshooter expressed interest. I make no claims about its validity, reliability, accuracy, etc. It was a nice thought experiment for me ;-)

 

[OldShooter]

... So this is just what I came up with in the last six hours since Oldshooter expressed interest. I make no claims about its validity, reliability, accuracy, etc. It was a nice thought experiment for me ;-)

I think that's very insightful. I often argue that talking about AA without knowing the overall financial condition and goals of the investor is a bad idea. You're really adding that concept to the idea of risk. Makes sense to me. DW and I could take a pretty significant $$ risk without there being any practical risk of ruining us financially. Others maybe not so much.

... Your posting style seems confrontational, seems to seek conflict and seems to propose words on behalf of others. It's OK, it doesn't bother me, my skin is pretty thick. ...

Sorry. I took the tone of your question as implying that you did not think quantifiable risk measures existed. I thought I was agreeing with you.

... I like standard deviation of return, max drawdown, downside capture ratio, value at risk and intra-portfolio correlation.

I dunno. To me all of those are just riffs on the theme of volatility as risk.

... Examples you gave fall into fraud and mismanagement in my view. What about meteors hitting the earth? Serious question. What about airliners flying into tall buildings in a global financial center? What about a global panic lasting nearly 2 years? ...

If I invest in a loser, I really don't care whether its demise was due to fraud, mismanagement, or its being taken out in some other way. Limiting the discussion to stocks, it would be very nice to have an algorithm that goes into alarm when it suspects fraud or mismanagement. It would also be interesting to see a study that tested whether measuring volatility was of any use in detecting stocks where investors lost a lot of money.

 

[chassis]

@OldShooter I think risk as the term has been used on this site is too broad.

I have listed risk measures I like. They are mainly related to equity portfolio construction.

An infinite array of other risk kpis can be developed from the likelihood of slipping on a bar of soap in the shower, to remaining life expectancy to anything else you can think of.

What quantified measure do you like for risk?

 

[SnowballCamper]

@OldShooter I think risk as the term has been used on this site is too broad.

I have listed risk measures I like. They are mainly related to equity portfolio construction.

An infinite array of other risk kpis can be developed from the likelihood of slipping on a bar of soap in the shower, to remaining life expectancy to anything else you can think of.

The question few ask or answer is, "Risk of what?" What adverse event do your KPIs measure the risk of? I looked them up on investopedia and such, but it isn't obvious to me.

 

[OldShooter]

... What quantified measure do you like for risk?

I don't know that there is one or even a suite of several. That's why Markowitz had to pick a poor one. Maybe there could be test battery including one tool where volatility was used in identifying high fliers nearly certain to crash. One time for Tesla I made up a back-of-envelope business scenario that assumed a $700 stock price, existing # of shares outstanding, reasonable product margins, and a reasonable P/E. That produced a sales volume number requiring that Tesla have 100% worldwide market share in cars and light trucks. Playing with Ben Graham fundamentals numbers like that, maybe a Monte Carlo approach, might result in a predictive tool or tools.

But it's pretty clear that no one has a magic tool to identify and avoid risky stocks, including Markowitz. If someone did, they would be consistent winners in the markets year after year. And we don’t see that. (Or maybe it is happening in secret. We can never rule that out.)

Absent a tool to identify risk, our best option is to diversify individual stock risk away and just accept market risk while riding on the long-term upwards market trend. IOW to ignore risk.

The question few ask or answer is, "Risk of what?" ...

I would say that risk is choosing an investment that has an above-50% probability of losing significant money over a buy-and-hold time period, years. That's a little simplistic vs formal risk/probability/cost-to-mitigate type risk analysis but I think it's useful. P/E is often used as a tool to identify this kind of risk.

 

[imnontrad]

If you have started "decumulation", just withdraw from each equity to bring it into balance again. No additional rebalancing needed.

 

[mrfeh]

If you have started "decumulation", just withdraw from each equity to bring it into balance again. No additional rebalancing needed.

Just withdrawing annual expenses may not be enough to get back to the target AA.

 

[imnontrad]

Just withdrawing annual expenses may not be enough to get back to the target AA.

then withdraw over a two year period.
OR consider the "three bucket" approach to retirement funds.

 

[mrfeh]

then withdraw over a two year period.
OR consider the "three bucket" approach to retirement funds.

I don't see how that solves the problem.

Cash is part of fixed income. If you withdraw more than your expenses, you are rebalancing.

 

[meleana]

I don’t. The only thing I did recently is redo my Traditional IRAs and Roth IRAs to make sure that mostly all the stock funds are in the Roth and mostly all the bond funds are on the Traditional.

 

[steelyman]

I don’t. The only thing I did recently is redo my Traditional IRAs and Roth IRAs to make sure that mostly all the stock funds are in the Roth and mostly all the bond funds are on the Traditional.

I’m similar but I have a secondary goal of achieving a specific split between Roth and traditional (I don’t want to do Roth conversions). It takes a while but I’m close.

 

[street]

The question few ask or answer is, "Risk of what?" What adverse event do your KPIs measure the risk of? I looked them up on investopedia and such, but it isn't obvious to me.

There is risk having to much cash, there is risk with to high of percent of equity. There is risk to being in the markets and there is a risk not being in the markets. I do agree "risk of what" and each person risk or what risk they feel is a risk to them.