Sequence Risk

[OldShooter]

Well that’s why I have a 50/50 AA, besides other volatility/end point characteristics, I also have plenty of room to both withdraw AND rebalance. I do have a lower $$ bound for fixed income.

Hmm ... Interesting. Essentially you are saying that you are keeping some equity money out of the market until the market goes down. Thus foregoing any upside on that money. Right?

 

[audreyh1]

I prefer to maintain a 50/50 AA, even if that means I miss out on the long term upside of a higher equity allocation. It has made it much easier for me to stay invested under a variety of market conditions, which for me is the important thing. The potential upside of a higher equity allocation isn’t important to me.

All of the studies I have studied and models I’ve run indicate my approach should meet my goals just fine.

 

[Gumby]

The FIREcalc spreadsheet shows the percentages and withdrawal amount based on the total portfolio value and the end results. But to actually implement it, you have to sell some assets to cover the withdrawal as well as selling and buying more assets to rebalance. It can be all calculated as a single step which is what I actually do, but assets will be sold based on how out of balance they are from your target, not based on the resulting AA.

Now some folks might rebalance and then withdraw proportionally, but it makes far more sense to do it all at once and calculate the rebalance after the withdrawal is removed, otherwise you’ll buy one of the asset classes and turn around and immediately sell it again to meet your withdrawal. Certainly not practical in a taxable account. I don’t think you can claim it’s proportional selling when you choose to rebalance before doing the withdrawal, because the rebalancing certainly wasn’t proportional.

As a practical matter, I think most of us would sell bonds to fund withdrawals when the stock market is down. And I know that people probably would do one net transaction, not two. I don't challenge you on that at all.

I'm mostly thinking about how you would construct the model in FIRECalc. For example, I think about a scenario in a year where the equities and stocks go up by the same exact percentage, such the the allocation remains rock solid at 60/40. Absolutely no re-balancing required. So where does the yearly draw then come from? You could actually choose any way you want, but what is the best a priori assumption for the model builder?

Similarly, what happens when you, for whatever reason, have chosen a 97/3 portfolio, you have a $1 million starting amount, the stock market drops 10% and you need to draw $40,000. You don't have the assets in fixed income to either re-balance or fund the yearly draw. Yes, it would be crazy to have that allocation, but from what I can see, FIRECalc will allow you to select up to 100% equities. How should the model address that?

If I were building the model and needed to accommodate such a wide range of possibilities, I would do one of 2 things:

1. Take the yearly draw from the equities/fixed in the proportion they have reached prior to re-balancing (e.g. 67%/33% for a nominal 60/40 portfolio) and then rebalance the remainder back to 60/40; or

2. Rebalance to 60/40 and then take the yearly draw 60% out of equities and 40% out of fixed. I think I showed above that the math comes out the same either way.

Otherwise, it would be way too complicated to accommodate the edge cases, only some of which I have listed above. Note that in either of my two methods the steps can be combined to only show a single net movement of money (as you note).


IF (and it's a big if) I am correct that the model builder takes the yearly draw in one of those two proportional methods, then FIRECalc incorporates sales of equities to fund draws during an equities bear market. Since most people will try to avoid doing that in real life, FIRECalc is probably conservative with respect to Sequence of Returns Risk.

Again, I'm not trying to say what you should do. I'm only trying to figure out how the model works. If anyone can shed light on this, I welcome it.

 

[Mountain skier]

I prefer to maintain a 50/50 AA, even if that means I miss out on the long term upside of a higher equity allocation. It has made it much easier for me to stay invested under a variety of market conditions, which for me is the important thing. The potential upside of a higher equity allocation isn’t important to me.

All of the studies I have studied and models I’ve run indicate my approach should meet my goals just fine.

I am 50/50 AA too. Audreyh1 do you mind sharing your approximate yearly expense amount and perhaps stash amounts, any pensions? I enjoy reading your posts. I wondering how similar your situation is to mine.

Thanks

 

[audreyh1]

I suspect FIREcalc calculates it all at once based on the spreadsheets it generates for our viewing pleasure.

And I have only one set of formulas in my spreadsheet that works in all cases - I don’t have any edge cases. It calculates how much to buy or sell from each fund based on the current amount in each fund (which depends on the current AA), the amount to withdraw, and the target AA.

It’s really not that hard and most calcs are based on total portfolio value. You have current portfolio value, portfolio value after withdrawal, and target value for each fund based on target AA of the portfolio value after withdrawal. Then you simply subtract the target value for each fund from the current value of each fund and it tells you how much to buy or sell for each fund. Then you pull your withdrawal from the core account and the resulting post-withdrawal portfolio is at target AA, i.e. rebalanced.

 

[FREE866]

I prefer to maintain a 50/50 AA, even if that means I miss out on the long term upside of a higher equity allocation. It has made it much easier for me to stay invested under a variety of market conditions, which for me is the important thing. The potential upside of a higher equity allocation isn’t important to me.

All of the studies I have studied and models I’ve run indicate my approach should meet my goals just fine.

question--
I just ran a 100% stock portfolio vs 50/50 stock /bond
20 years
$1,000,000 start
$50,000 withdrawn annually
the 100% stock had an avg final balance over 50% higher and success rate was 92% vs 90.4% vs the 50/50


not sure how that is possible, but that tells me 100% stock is way to go no?

 

[audreyh1]

It depends on whether you can live with the annual volatility of 100% stocks. Generally over 30 years at 4% 50/50 has better survival characteristics than 100% stocks.

I don’t have much use for a larger ending portfolio, as I’ll be ancient or dead.

You have to be clear on your goals.

 

[FREE866]

It depends on whether you can live with the annual volatility of 100% stocks. Generally over 30 years at 4% 50/50 has better survival characteristics than 100% stocks.

I don’t have much use for a larger ending portfolio, as I’ll be ancient or dead.

You have to be clear on your goals.

I hear you. I was just surprised by the % success.

 

[Navigator]

It depends on whether you can live with the annual volatility of 100% stocks. Generally over 30 years at 4% 50/50 has better survival characteristics than 100% stocks.

I don’t have much use for a larger ending portfolio, as I’ll be ancient or dead.

You have to be clear on your goals.

I am not advocating for 100% stock portfolios, but I would like to point out that one needs to specify a withdrawal rate when considering survivability. For withdrawal rates above 5%, a well balanced 100% stock portfolio actually has better survivability rates than a 50/50 portfolio. Such withdrawal rates could be necessary for those who do not have large savings or could happen natural if the marked crashed immediately after someone retired.

 

[Navigator]

I also think people overstate the value of a flexible budget.

The whole point of a conservative WR is that it will survive those downturns without cutting spending.

-ERD50

Actually, I think setting a low withdrawal rate just means one permanently cuts spending before they get started. Some people are wealthy enough So that this does not matter, and I certainly congratulate anyone in this position. Many of us, however, need to make trade offs between when we retire and retirement lifestyle. Setting an artificially low withdrawal rate means either working longer or not having as much money to spend as we would like.

 

[ERD50]

...
I'm mostly thinking about how you would construct the model in FIRECalc. ...

Similarly, what happens when you, for whatever reason, have chosen a 97/3 portfolio, you have a $1 million starting amount, the stock market drops 10% and you need to draw $40,000. You don't have the assets in fixed income to either re-balance or fund the yearly draw. Yes, it would be crazy to have that allocation, but from what I can see, FIRECalc will allow you to select up to 100% equities. How should the model address that?

If I were building the model and needed to accommodate such a wide range of possibilities, I would do one of 2 things:

1. Take the yearly draw from the equities/fixed in the proportion they have reached prior to re-balancing (e.g. 67%/33% for a nominal 60/40 portfolio) and then rebalance the remainder back to 60/40; or

2. Rebalance to 60/40 and then take the yearly draw 60% out of equities and 40% out of fixed. I think I showed above that the math comes out the same either way.

Otherwise, it would be way too complicated to accommodate the edge cases, only some of which I have listed above. Note that in either of my two methods the steps can be combined to only show a single net movement of money (as you note). ...

I'm only trying to figure out how the model works. If anyone can shed light on this, I welcome it.

You are way over-complicating this. There are no edge cases or decisions, it is simple arithmetic.

But I will say that my statement "you won't sell equities in a downturn" was an over-simplification. I should have said "in an extreme downturn". With the example $40,000 withdrawal, if less than $40,000 is needed to re-balance, then yes, you'll still be selling some equities. You observed this in your "Edit to add" example at the - 9.5% point.

The simple way to look at it is:

A) Take your TOTAL EOY Balance.

B) Subtract your withdrawal $ amount from that.

C) Calculate your new $ allocations ( based on Total Balance after W/D from "B"). Done.

FIRECalc doesn't really need to deal with selling or buying, those new starting $ amounts for equity and fixed is all it needs. IRL, if we were strict re-balancers, we would make our withdrawal, plus any amounts to buy/sell to re-balance.

Even your 97/3 AA doesn't cause any issues for the simple arithmetic. (I'll try posting the spreadsheet clips next).

A couple more "IRL" points:

A) I know there are some strict re-balancers here, and that's fine. But I suspect there are many who re-balance when they get "out of whack" by enough to concern them (5%?). And they probably take a more simplistic (pragmatic?) approach and take from whichever asset class is higher than their target, and don't try to split hairs. So if your 75/25 target is now 80/20, take your W/D from equities, don't bother with the math, just get closer. And vice-versa. One transaction.

B) IRL, we are probably taking our dividends as they accrue. For an ~ 4% WR, divs probably make up about half of that, so the WD amount is half as much, so therefore less to sell from stocks in a mild downturn, and less overall effect from this W/D.

Back to FIRECalc, I' pretty sure the data sources they use include re-investment of divs, so the program really has no choice but to use the entire W/D amount in its EOY W/D & re-balance calculations.

-ERD50

 

[ERD50]

Actually, I think setting a low withdrawal rate just means one permanently cuts spending before they get started. Some people are wealthy enough So that this does not matter, and I certainly congratulate anyone in this position. Many of us, however, need to make trade offs between when we retire and retirement lifestyle. Setting an artificially low withdrawal rate means either working longer or not having as much money to spend as we would like.

True. You place your bets, you take your chances.

There's a place for both, but I stand by my point that I doubt that many have really considered how deep a cut they have to make to make a real difference, or how they decide when to do it. As I said, we can have a 2-3 year downturn - cutting a bunch doesn't really make much difference over such a short time, but it could mean a lot to your quality of life.

My personal decision was to be conservative enough to not worry about short term downturns. My regular budget doesn't contain a lot that I'd be OK with cutting, I don't live extravagantly by most standards, but to live comfortably in the way I like is still not cheap. Yes, it means I had to save more. It's a trade-off, no right/wrong answer.

-ERD50

 

[ERD50]

Here's a few screenshots of the spreadsheet mentioned below. One clip of two of the examples in this thread, one with the formulas exposed. Entries are made in the cyan shaded cells.

The formulas could be done in one step, but it would be harder to follow. Very, very simple in a program.

(edit - missed a row on the original screen shot example, this should fix it)

-ERD50

 

[Gumby]

I fear I have not been clear on this. I am not talking about figuring out your own draw and rebalance amounts every year. That is mechanistic and can vary as you please to do it (strict or loose rebalancing or dividends or no). What I am talking about is the predictive role of FIRECalc - the way that it uses historical return data and your particular hypothetical portfolio amount and asset mix to calculate ultimate success rates. To do that, the model must have an algorithm for taking the draw one year so that it can establish the precise asset mix for applying the next year's return. The easiest rule, I think, is to take from equities and fixed income in the proportion to which they have grown prior to re-balancing.

Why do I find this important - because in real life, people are apt to actually fund their draw mostly if not entirely from fixed income when the stock market is down. If FIRECalc assumes that you will fund 60/40 no matter what, it is more conservative with respect to the Sequence of Return Risk.

ETA: One last try. The OP expressed some hesitation about a 100% result from FIRECalc given a potential adverse Sequence of Returns in the first 2 years. You, ERD50, pointed out that he need not be so concerned because he could just not sell equities into a down market. I added that if you strictly followed FIRECalc, it would in some circumstances require that he sell equities in a down market (which we subsequently established to be true). So, FIRECalc is conservative. If it still gives you 100% even selling equities into an 8% market drop, then you are even more assured of success because you are not going to sell those equities. You are most likely going to dip into your fixed income to ride out the bear. I think you and I are in agreement. But then Audrey said that it is not correct that FIRECalc assumes some sort of proportional withdrawal from fixed/equities. I think it must.

 

[RunningBum]

True. You place your bets, you take your chances.

There's a place for both, but I stand by my point that I doubt that many have really considered how deep a cut they have to make to make a real difference, or how they decide when to do it. As I said, we can have a 2-3 year downturn - cutting a bunch doesn't really make much difference over such a short time, but it could mean a lot to your quality of life.

My personal decision was to be conservative enough to not worry about short term downturns. My regular budget doesn't contain a lot that I'd be OK with cutting, I don't live extravagantly by most standards, but to live comfortably in the way I like is still not cheap. Yes, it means I had to save more. It's a trade-off, no right/wrong answer.

-ERD50

This is why I like VPW. You start making small changes in a downturn. If it recovers, you can resume normal spending. If it continues down, you keep cutting as needed. There is no ideal way to keep it from hurting if the downturn is severe, but at least you aren't faced with the decision of when to abandon your plan and make deep cuts. VPW guides you to that.

 

[ERD50]

....

Why do I find this important - because in real life, people are apt to actually fund their draw mostly if not entirely from fixed income when the stock market is down. If FIRECalc assumes that you will fund 60/40 no matter what, it is more conservative with respect to the Sequence of Return Risk.

OK, so we know (or at least are fairly certain) that FIRECalc does a mechanical re-balance to the $1.00 at the start of each year.

You seem to be questioning if that mechanical adjustment provides steeper drops in a downturn than what most of would do with our approximations and real-life withdrawals?

Hard to say w/o defining some 'real life' algorithms. And there are so many possible scenarios. Let's say we had four annual ~ 10% declines in a row, and our equities are down ~ 40%. FIRECalc might be withdrawing from stocks in each of those years (depending on specifics, but let's go with that). Say bonds are returning 3%.

Maybe a forum member would see that the first year, their 75/25 moved to 72.4/27.6, so OK, pull all from fixed since stocks dropped. That gets them to 75.6/24.3 - hair splitting. And remember, we are really only w/d re-bal with about half as much, since we are not reinvesting divs.

So the next year, or the next, we might swing a little the other way. W/o going into 10 iterations of this, I'd say it looks like it would have minimal effect. And it might be a wash on the way back up. Heck, from the data I've seen, re-balancing alone is pretty much a wash, so I would expect modulating it a bit would really be near zero effect?

-ERD50

 

[Gumby]

OK, so we know (or at least are fairly certain) that FIRECalc does a mechanical re-balance to the $1.00 at the start of each year.

You seem to be questioning if that mechanical adjustment provides steeper drops in a downturn than what most of would do with our approximations and real-life withdrawals?

Hard to say w/o defining some 'real life' algorithms. And there are so many possible scenarios. Let's say we had four annual ~ 10% declines in a row, and our equities are down ~ 40%. FIRECalc might be withdrawing from stocks in each of those years (depending on specifics, but let's go with that). Say bonds are returning 3%.

Maybe a forum member would see that the first year, their 75/25 moved to 72.4/27.6, so OK, pull all from fixed since stocks dropped. That gets them to 75.6/24.3 - hair splitting. And remember, we are really only w/d re-bal with about half as much, since we are not reinvesting divs.

So the next year, or the next, we might swing a little the other way. W/o going into 10 iterations of this, I'd say it looks like it would have minimal effect. And it might be a wash on the way back up. Heck, from the data I've seen, re-balancing alone is pretty much a wash, so I would expect modulating it a bit would really be near zero effect?

-ERD50

I was editing while you were posting, so I'll just repeat it here:

ETA: One last try. The OP expressed some hesitation about a 100% result from FIRECalc given a potential adverse Sequence of Returns in the first 2 years. You, ERD50, pointed out that he need not be so concerned because he could just not sell equities into a down market. I added that if you strictly followed FIRECalc, it would in some circumstances require that he sell equities in a down market (which we subsequently established to be true). So, FIRECalc is conservative. If it still gives you 100% even selling equities into an 8% market drop, then you are even more assured of success because you are not going to sell those equities. You are most likely going to dip into your fixed income to ride out the bear. I think you and I are in agreement. But then Audrey said that it is not correct that FIRECalc assumes some sort of proportional withdrawal from fixed/equities. I think it must.

 

[ERD50]

OK, so I ran my spreadsheet out for 4 years. And yes, it's going to be very scenario dependent.

So stocks are going down, so you think you are genius by selling from bonds, so you don't sell stocks low? Fine, but if the downturn continues for another year or two, that means you have more stocks exposed to those further downturns. After 4 years of selling bonds (assuming -10% for stocks, +3% for bonds), the straight arithmetic portfolio was about 0.9% ahead of the human who took the whole W/D from bonds the 1st 3 years, and then the whole W/D from stocks the 4th year (the fourth year was the first time they were above 75%, after 3 years of pulling bonds only).

And that's with pulling the full 4%, which would be more like 2% if divs make up 2%.

I think this reaffirms to me that our portfolios are not very sensitive to AA, and they aren't very sensitive to re-balancing. And any small difference varies with the scenario - back to Crystal Ball gazing, I think.

ETA: this also cross posts with your post # 42. I'll stop now so it can sink in for each of us. But I think when you read the above, you might agree that it's uncertain whether FIRECalc would be over or understating the situation?

-ERD50

 

[Gumby]

OK, so I ran my spreadsheet out for 4 years. And yes, it's going to be very scenario dependent.

So stocks are going down, so you think you are genius by selling from bonds, so you don;t sell stocks low? Fine, but if the downturn continues for another year or two, that means you have more stocks exposed to those further downturns. After 4 years of selling bonds (assuming -10% for stocks, +3% for bonds), the straight arithmetic portfolio was about 0.9% ahead of the human who took the whole W/D from bonds the 1st 3 years, and then the whole W/D from stocks the 4th year (the first year they were above 75%, after 3 years of pulling bonds only).

And that's with pulling the full 4%, which would be more like 2% if divs make up 2%.

I think this reaffirms to me that our portfolios are not very sensitive to AA, and they aren't very sensitive to re-balancing. And any small difference varies with the scenario - back to Crystal Ball gazing, I think.

-ERD50

I think that is as well as we are going to do today. It does appear to be quite scenario dependent. In any event, it has been a pleasure tossing this idea around with you.

 

[audreyh1]

This is why I like VPW. You start making small changes in a downturn. If it recovers, you can resume normal spending. If it continues down, you keep cutting as needed. There is no ideal way to keep it from hurting if the downturn is severe, but at least you aren't faced with the decision of when to abandon your plan and make deep cuts. VPW guides you to that.

Pretty much how the more basic %remaining portfolio works too. And if the portfolio grows faster than inflation early on you get to take advantage of that too.

And having plenty of discretionary spending and thus flexibility means you can more easily absorb the income variability.

That’s the approach I’ve chosen. I prefer dealing with varying income to having the portfolio go to zero.

 

[SecondCor521]

I was editing while you were posting, so I'll just repeat it here:

ETA: One last try. The OP expressed some hesitation about a 100% result from FIRECalc given a potential adverse Sequence of Returns in the first 2 years. You, ERD50, pointed out that he need not be so concerned because he could just not sell equities into a down market. I added that if you strictly followed FIRECalc, it would in some circumstances require that he sell equities in a down market (which we subsequently established to be true). So, FIRECalc is conservative. If it still gives you 100% even selling equities into an 8% market drop, then you are even more assured of success because you are not going to sell those equities. You are most likely going to dip into your fixed income to ride out the bear. I think you and I are in agreement. But then Audrey said that it is not correct that FIRECalc assumes some sort of proportional withdrawal from fixed/equities. I think it must.

I'll probably not make anything better, but I'll try anyway. I'm an optimist.

The way a person behaves is up to them. Personally I try to maintain my AA all the time; I wouldn't try to second-guess a down market.

The way FIREcalc actually works is every year calculating the 12/31 balance based on the AA and the historical performance figures for the asset classes during that year. It then subtracts the withdrawal amount. It then rebalances back to the AA, completely ignoring any tax effects (that's on us to account for). That then is the situation on 1/1 of the next year. Then that annual process repeats for all 30 (or whatever) years of the run.

As some other poster has noted above, whether one rebalances first then proportionately withdraws, or withdraws then rebalances, it's all the same (excepting possibly the number of transactions, which I view as irrelevant).

I'm pretty sure that's all spelled out on the FIREcalc pages somewhere.

 

[Hopeful]

Better it should be just preceding where you still have the flexibility to potentially wait longer to retire, and save more, rather than it happen just after when you begin spending down your investments into the downturn.

That is good point, and in the back of my mind I keep wishing the correction would come AND go sooner than later. While it is nice to have the option of continuing to w***, the thought of having to delay RE is what would be disappointing.