Risk Perspective (I'm Probably Confused)

[RetireAge50]

See Table Below.
The Expected Income (a portfolio could generate) after a 50% stock market crash seems to be about the same regardless of your asset allocation.
Would it not make sense to stay mostly in stocks for the higher expected income and then during down markets reduce your income to the lower levels until it recovers?

Chart Explanations:
Allocation: Stocks/Bonds
Historical Real Return: Used 7% Stocks 2% Bonds
Expected Income: Excel Function PMT(Rate,40,$1,000,000)
After Crash: Same Excel Formula ($1M Reduced Based on Stock %)

 

[samclem]

Some people cannot take a 50% reduction in their monthly checks--their spending needs aren't that flexible, so they go with a portfolio that has less volatility (and they give up some expected return).

As I'm sure you know, markets don't generate constant yearly return values. The ups-and-downs, and the rebalancing during those times, allows a retiree to "buy low and sell high" if the portfolio has both stocks and bonds and is rebalanced regularly. This reduces the volatility in the portfolio
s value and can even improve overall returns compared to a 100% equity portfolio. Now, it doesn't always happen that stocks and bonds move independently of each other, but it happens frequently.

The withdrawal calculation (% of year end value vs % of starting value adjusted annually for inflation) has a big impact on portfolio survivability. If you can be satisfied to take a considerable "hit" in your annual withdrawals when the market is down (thereby selling fewer shares at those times), the portfolio has a much better chance of recovering when the market recovers. This is true regardless of the allocations, but is more significant if the allocations to stock are high.

 

[galeno]

There are two risks to a retirement portfolio: shallow (volatility) and deep (inflation). Balance between the two = success in retirement.


When finance people talk about "risk tolerance" they are talking about shallow risk. Many retirees cannot tolerate the volatility of an equity heavy portfolio.


During a bear market many will panic and do stupid things like selling all stocks at the bottom and then miss the ride back up.


On the other extreme there are people who are so afraid of volatility that they try to use only bonds. These retirees run the risk of running out of money before they run out of life.


For me, the "sweet spot" for equities during a 30+ year retirement = 40-60%.

 

[RetireAge50]

In this example a 70/30 allocation would provide an expected income of $62,320 which would drop to $40,508 after the crash.

If a person wanted to avoid such an income drop they might decide to go with a 30/70 allocation which would provide an expected income of $46,827 which would drop to $39,803.

So to me it seems like taking a higher stock allocation would provide about the same base income and provide higher incomes for non-crash periods.

 

[samclem]

In this example a 70/30 allocation would provide an expected income of $62,320 which would drop to $40,508 after the crash.

If a person wanted to avoid such an income drop they might decide to go with a 30/70 allocation which would provide an expected income of $46,827 which would drop to $39,803.

So to me it seems like taking a higher stock allocation and budgeting the lower amount would be better.

Subsequent edit--I figured out your setup)
[-]Sorry, I don't have time now to reverse-engineer your calculations. Are you stipulating that the 30s/70b portfolio and the 70s/30b portfolio suffered the same 50% loss in value? If so--that's the problem with your setup. People hold more bonds to reduce the size of that drop, and it usually works.
[/-]
FIRECalc can do a good job of showing you how portfolios of various composition would have done over time. Set it up to withdraw a % of the year-end value and you'll see how high stock and high bond allocations have done (and the variability of the income they have provided, in terms of inflation-adjusted spending available) over many, many actual market cycles using real data. It will be more useful to you than this spreadsheet, I think.

 

[RetireAge50]

Only the stock portion for each allocation loses 50%. So the 30/70 portfolio would lose 15% (50% of 30%).

I do use firecalc and it is a good downside tool. But I am not sure it fully addresses what an appropriate expected income would be.

 

[haha]

In this example a 70/30 allocation would provide an expected income of $62,320 which would drop to $40,508 after the crash.

If a person wanted to avoid such an income drop they might decide to go with a 30/70 allocation which would provide an expected income of $46,827 which would drop to $39,803.

So to me it seems like taking a higher stock allocation would provide about the same base income and provide higher incomes for non-crash periods.

This must be some non-standard sense of income. No way is a 70:30 portfolio, or a 100:00 portfolio or any other diversified portfolio going to provide that kind of pre crash income, with todays interest rates and S&P dividend yields. Maybe they mean some sort of supposed total return?

Ha

 

[Independent]

See Table Below.
The Expected Income (a portfolio could generate) after a 50% stock market crash seems to be about the same regardless of your asset allocation.
Would it not make sense to stay mostly in stocks for the higher expected income and then during down markets reduce your income to the lower levels until it recovers?

As a general rule, people who have lots of flexibility in their spending can invest in higher risk/reward classes. I think you've demonstrated why they might want to do that.

The problem is people who have non-flexible spending plans, but base their retirement date and spending on the "expected" returns of high risk/reward classes.

As Sam suggested, if you're thinking about this for yourself, try looking at how the various allocations really played out over historic periods (and I'd add, in Monte Carlo simulations). Pay close attention to the first few years, as "a few years" is often our emotional time horizon.

 

[Animorph]

This is a pretty simple model. FIRECalc shows best portfolio survivability in a broad range around 50% stocks, like the right hand column. And 100% stocks comes very close, maybe even closer than the table shows. And 100% stocks gives you the potential of a higher portfolio value near the end. So I do stay 100% equities for my AA. I try to raise cash for expenses early if possible, as soon as the portfolio reaches a projected beginning of the year value. Hopefully that leaves me with cash in a downturn without being a drag on performance. Or no big bear markets would be nice too.

 

[Animorph]

This must be some non-standard sense of income. No way is a 70:30 portfolio, or a 100:00 portfolio or any other diversified portfolio going to provide that kind of pre crash income, with todays interest rates and S&P dividend yields. Maybe they mean some sort of supposed total return?

Ha

Looks like he's just calculating annuity amounts for 40 years, so it's a perfect world other than the crash.

 

[RetireAge50]

Ha. When I say income I meant portfolio withdrawals.

Independent. Understand and agree with your comment.

I'm thinking people should budget the downside but actually withdraw based on actual and perhaps invest for the higher withdrawals knowing the downside still meets their budget.

 

[daylatedollarshort]

A different way of looking at retirement assets and income is here:

Matching strategy - Bogleheads

DH and I understand that there is a high probability we could make more money over the long term holding more stocks. But the pain of losing half of our net worth at our ages would be severe while the gain of earning more would not make that much difference in our lifestyles, since we are both cheap dates and we've gotten our basic expenses pretty low these days. And at our ages maybe the long term would be longer than our remaining life spans. So we stick more with the kinds of investments in the liability matching strategy in the Bogleheads wiki.

 

[haha]

Ha. When I say income I meant portfolio withdrawals. ��

Independent. Understand and agree with your comment.

I'm thinking people should budget the downside but actually withdraw based on actual and perhaps invest for the higher withdrawals knowing the downside still meets their budget.

Thank you.

Ha

 

[samclem]

I do use firecalc and it is a good downside tool. But I am not sure it fully addresses what an appropriate expected income would be.

Okay, everybody sees and learns things a bit differently. FWIW, here are some depictions of interest. Assumption: $750K starting value, withdraw 4% of the end-of-year balance each year. The graphs show the actual real value of withdrawals (starting-year dollars). The text boxes address the value of the portfolio itself.

Here's the graph for 100% equities.
(see "100S.png" attachment

Here is how your portfolio would have fared in each of the 114 cycles. The lowest and highest portfolio balance throughout your retirement was $514,458 to $3,493,756, with an average of $1,462,771. (Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

Here's the 50s/50b output:
(see "50-50.png" attachment)

Here is how your portfolio would have fared in each of the 114 cycles. The lowest and highest portfolio balance throughout your retirement was $420,843 to $1,828,852, with an average of $834,382. (Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

Here's 100% bonds (5 year Treasuries):
(See "0S.png" attachment)

Here is how your portfolio would have fared in each of the 114 cycles. The lowest and highest portfolio balance throughout your retirement was $182,321 to $1,214,530, with an average of $415,388. (Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

Takeaways:
- Stock prices sometimes don't just drop in a one-time plunge, but they can descend and then stagnate at low levels. So, if we look at the first 8 years of each depiction above and we especially look at how many of each lines approach the blue "poverty level" line. Over that first 8 years, this happens much more frequently for the 100% stock portfolio than the 50/50 or the 100% bonds portfolio. This is the "shallow" risk (volatilty) that galeno mentioned, and it's what people try to avoid by having more bonds. The construct of your sample spreadsheet doesn't do a good job of showing this risk, I think FIRECalc helped me understand it better.

- Now, look at all three charts at 20-30 years. Many of the lines on the 100% bonds chart have dropped below the poverty level, and none of the lines for the 50/50 or 100% stock portfolio have done so. The 50/50 and 100% stock portfolios help avoid this deeper risk--long term failure to keep up with inflation.

- Other: Be sure to observe the scales on the Y axis--they vary a lot across the three depictions.

So, this is why people choose balanced portfolios. They are the "sweet spot" and minimize both shallow risk (volatility) and deep risk (losing value to inflation). FIRECalc (and other models) can do a lot to help illuminate these issues, within their limits.

Good luck

 

[RetireAge50]

Thanks for running the charts and explaining the shallow and deep risk. Makes sense.

I wish Firecalc could adjust spending EACH year using the remaining balance, time remaining, and historical average rates. Using the 4% times the balance does not amortize the portfolio down to zero.

 

[samclem]

I wish Firecalc could adjust spending EACH year using the remaining balance, time remaining, and historical average rates. Using the 4% times the balance does not amortize the portfolio down to zero.

I don't know that there has been much demand for it, since the "time remaining" is unknown to most of us and so we avoid trying to run the portfolio to zero on a particular date.

If you don't want a fixed % of the remaining balance each year, and you don't want a starting balance adjusted for inflation each year, maybe you would like the Variable Percentage Withdrawal (VPW) method. It has some fans, and sounds like what you are after. Here's a long thread on it. Like the standard RMD calculation, the withdrawal rate % is recalculated each year based on your life expectancy. It's a bit more aggressive than I'm comfortable with, but I will run it anyway as another data point as I withdraw funds.
Edited to add: The link to the thread above still works and I think the discussion on VPW is very much worth reading, but some of the internal links there no longer work. For a "quick grip" on VPW, a good place to start is the Bogleheads wiki page on it. Here