Kitces: Preparing for Lower Long-Term Returns

[NW-Bound]

...No matter which method a retiree uses, occasional "reality checks" will be needed to assure things aren't headed to the sky, or the toilet.

It's the cases for the latter (the toilet) that I worry about, and not the first (the blue sky).

Think about it, if one retires with $1M and initially drawing $40K, will he not be "ratcheting" it up when his portfolio will have doubled (in real terms no less), and he is 10 or 15 years older and closer to that 6-ft deep hole? That would be the least problem for anybody. I would not worry about dying rich. It's dying young that would bother me more.

On the other hand, having to live on 1/2 of what he started out is going to be a lot tougher. In my case, I think I can handle that 50% reduction with the aid of SS plus the Bernicke effect. And being on Medicare should help, compared to what I am paying now.

 

[audreyh1]

Right, but:
- This is 4.5%, not the "standard" 4%. So, a "heftier" starting WR than most people count on
- The vast majority of portfolios stayed fairly close to the starting value, and annual withdrawals stayed generally close to the central starting value. Things are a lot different if instead we use "starting amount adjusted for inflation" methods.
- High side: Only about 5 of the 116 runs "exploded", with ending values double the size of the starting value (leaving tons of money unspent by the retiree). In general, portfolios that did well returned the $$ to the retiree as withdrawals. Using a fixed 4.5% of starting value adjusted for inflation method, about 15 portfolios ended up "exploding (ending value at least twice as high as starting value).
- Low side: Some of the portfolios (15%?) had long periods of providing about 1/2 of the starting withdrawals for extended periods. I'd bet these are the same ones that would have crashed entirely (zero balance) if we'd used a "4.5% starting withdrawal adjusted for inflation" withdrawal method. That produced a 23% failure rate.

No matter which method a retiree uses, occasional "reality checks" will be needed to assure things aren't headed to the sky, or the toilet.

Thanks for the details. I can't seem to read the .xls files generated by FIRECALC (other than the text), so it's good to get a better feel for the results.

 

[audreyh1]

I did some FIRECALC runs today, and it reinforces my instincts to allow unspent funds to accumulate. Even on cases with lower withdrawals such as 3%, 3.25%, 3.3% a few runs seem to drop to about 55% of starting value, inflation adjusted, before recovering up to around where you started, and a more down to 64% of starting value. So you might face a year, or a few years together, where income has dropped almost in half.

 

[samclem]

I did some FIRECALC runs today, and it reinforces my instincts to allow unspent funds to accumulate. Even on cases with lower withdrawals such as 3%, 3.25%, 3.3% a few runs seem to drop to about 55% of starting value, inflation adjusted, before recovering up to around where you started, and a more down to 64% of starting value. So you might face a year, or a few years together, where income has dropped almost in half.

And, since usually those dips don't last very long (stocks >generally< recover within about 3 years IIRC, esp with reinvested dividends), it wouldn't take a huge pile money to ease through a rough patch. And, if a big expense comes up (new roof, emergency car purchase, etc) when stocks are down, then having a slug of cash would be useful.

Another option for those who don't like setting aside the cash like this would be to look at the results of using the Clyatt 95% rule (which FIRECalc will model easily). It avoids year-to-year withdrawal changes of more than -5% (plus inflation), has a negligible impact on portfolio survival, and avoids the cash "drag."

 

[NW-Bound]

Those squiggly lines bunch together, and that makes it difficult to see how long a bad period has lasted in the past.

I tried to use the spreadsheet output option, but the data that is downloaded and imported into my PC is not valid. I have done this in the past, but this time the data does not make sense and does not agree with the "squiggly line" chart.

 

[audreyh1]

Those squiggly lines bunch together, and that makes it difficult to see how long a bad period has lasted in the past.

I tried to use the spreadsheet output option, but the data that is downloaded and imported into my PC is not valid. I have done this in the past, but this time the data does not make sense and does not agree with the "squiggly line" chart.

Try renaming the files from .xls to .html and open it with a browser. You can then select and paste the data into a spreadsheet.

For the second file, the formulas in the second half won't work, but the numbers in the first half will display properly. You can at least see the numbers and you could past the following formulas into a spreadsheet if you really want to, just make sure you have the proper alignment.

 

[NW-Bound]

I was able to import them. The data does not make sense.

 

[audreyh1]

The first file seems to be the dry run data 30 years for each year 1871 to 2015. But it looks like it's nominal $.

The second file is a single run from 1960 to 1989 in real $. I assume you can pick the year to do your run so you can test the worst cases in the first file?

 

[audreyh1]

And, since usually those dips don't last very long (stocks >generally< recover within about 3 years IIRC, esp with reinvested dividends), it wouldn't take a huge pile money to ease through a rough patch. And, if a big expense comes up (new roof, emergency car purchase, etc) when stocks are down, then having a slug of cash would be useful.

Another option for those who don't like setting aside the cash like this would be to look at the results of using the Clyatt 95% rule (which FIRECalc will model easily). It avoids year-to-year withdrawal changes of more than -5% (plus inflation), has a negligible impact on portfolio survival, and avoids the cash "drag."

Right - you don't have to accumulate a large pile. In fact if you have an extra year's expenses set aside, that would probably bridge you across a three year run of significantly lowered income with just a slight belt tightening if necessary.

There are definitely other things you might want to have some additional cash sitting around for - an unexpected large medical expense, for example.

 

[NW-Bound]

The first file seems to be the dry run data 30 years for each year 1871 to 2015. But it looks like it's nominal $.

True, it does not look like it is inflation-adjusted, hence not too useful.

The second file is a single run from 1960 to 1989 in real $. I assume you can pick the year to do your run so you can test the worst cases in the first file?

I've got a lot of error in this file. Perhaps both of my MS Excel and OpenOffice copies are outdated.

 

[NW-Bound]

Right - you don't have to accumulate a large pile. In fact if you have an extra year's expenses set aside, that would probably bridge you across a three year run of significantly lowered income with just a slight belt tightening if necessary...

Although I have always been one who has a lot of cash on hand (never less than 20%), I wonder how useful this cash would be in a prolonged stagflation period like in the past. The recent market crashes of 2003 and 2008 were quite sharp like a V shape, but the wide U shape was perhaps more frequent (while you live in that U shape, it feels like an L).

When inflation ran in the double digit like the late 70s for several years in a row, cash got hammered badly. You would need gold!

 

[NW-Bound]

...When inflation ran in the double digit like the late 70s for several years in a row, cash got hammered badly. You would need gold!

... or I-bonds.

I have been keeping several years of expenses in I-bonds, and it hurts my overall portfolio return. But it will at least keep the purchasing power. Gold on the other hand will even gain in value during inflationary periods. I recall in 1980 gold price reached $900. And that's 1980 dollars, which would be $2700 now. Gold is now $1200/oz.

See, if we can tell in advance what is going to happen, we know exactly what to do.

 

[audreyh1]

When inflation ran in the double digit like the late 70s for several years in a row, cash got hammered badly. You would need gold!

Well, bonds got hammered badly. Cash, if you invested in shorter-term CDs and such, not so much.

 

[samclem]

... or I-bonds.

I have been keeping several years of expenses in I-bonds, and it hurts my overall portfolio return. But it will at least keep the purchasing power. Gold on the other hand will even gain in value during inflationary periods. I recall in 1980 gold price reached $900. And that's 1980 dollars, which would be $2700 now. Gold is now $1200/oz.

See, if we can tell in advance what is going to happen, we know exactly what to do.

We've got some of the "great" I-Bonds purchased when they offered a 3+% real interest rate. The "problem" is that I recognize now (in retrospect) that they are such a great deal that I'd be >very< reluctant to cash them out prior to maturity just to get through a period of low returns. While they do contribute to our net worth, they are, effectively, off the table for selling. OTOH, I'd have no difficulty selling gold if the price was up and our stocks were down.

 

[MooreBonds]

We've got some of the "great" I-Bonds purchased when they offered a 3+% real interest rate. The "problem" is that I recognize now (in retrospect) that they are such a great deal that I'd be >very< reluctant to cash them out prior to maturity just to get through a period of low returns. While they do contribute to our net worth, they are, effectively, off the table for selling. OTOH, I'd have no difficulty selling gold if the price was up and our stocks were down.

LOL - I have the same "problem" as you. I always considered my I-bonds (about 5% of current net worth) with an average real rate of 3.3% to be my emergency fund...but now they have become a bit quasi-sacrosanct due to their relative high yield. The benefit is that I'm still working. But if I were retired, depending on how much the portfolio value and dividends dropped, I'd likely just live off of the dividends from my portfolio - which is pretty much my goal anyway when I do FIRE. If we hit a prolonged period of true disaster, then the I-bonds would slowly be taken to the chopping block to supplement the dividends. Any other alternatives like a home equity loan would likely incur a net interest expense that would be less than the benefit of keeping the I-bond.

Unless I could get a negative interest rate Home Equity loan! (which is looking less and less likely, but which did exist for a short period of time in some European countries).

 

[NW-Bound]

I have thought about this some more, and rediscovered what I saw before (and kept forgetting despite my self-proclaimed "superior" memory). That is there are periods where there was just no gain to be had no matter where you look. Cash or CD did not save you, because, well everything was like cash: all showed zero return!

Using data from the Federal Reserve in St. Louis, I found the following for the period of 1/1966 to 1/1983, a period of 17 years.

$1 invested in stocks would become $3.04.
$1 invested in 3-month T-bills would become $3.175.
$1 invested in 10-year constant maturity T-bond would become $2.34.

However, the cumulative inflation factor over this 17-year period is 3.08.

So, your initial $1 would look like this after 17 years with inflation adjustment:

Stock: $0.99
T-Bill: $1.03
T-Bond: $0.76

There's no point in keeping a lot of cash (CD)! Everything looks like CD (or slightly worse), when averaged out over this 17-year period.

Now, that makes everything easy. If your WR is 3%, then after 17 years you will have spent about 1/2 of your stash, and have 1/2 of it left. If your WR is 4%, then you have about 1/3 left. It would be less if you hold long bonds. Don't hold long bonds!

It's not the end of the world. Your portfolio probably will survive for another 13 years to make it a 30-year retirement. However, you do not have to worry about dying rich.

 

[NW-Bound]

Now, one can go back to read the transcript of the interview with Kitces in the OP.

He said: "this is just the challenge of being in a low-return environment. Having low yields arguably is one of the things that's actually let equity valuations get as high as they are. But all that really means is, good news, you can get 2% in dividends on your stocks, which looks nice compared with getting zero or 1% in your short-term fixed income. But we're still choosing between dividends of 2% and fixed-income yields of less than 2%. This is still kind of like asking, "Do you want your low returns or do you want your low returns?"

It's not just Shiller and his PE10, but every pundit including Bogle has warned us about low return of both stocks and bonds in the years ahead. Of the PE10, Kitces said "we find out that Shiller P/E ratios are actually quite good at predicting things like 15-year real returns in equities".

Look at the PE10 for yourself, and see what it was in 1966 and what it is now. See how it decreased during the period of 1966-1983 that I described. The stock market was flat after inflation, even though businesses were doing OK. That's the effect of PE contraction. And everybody, Bogle included, said that PE contraction is something one must allow for.

See: Shiller PE Ratio

PS. One can easily see long periods of no growth using FIRECalc. Set the WR to 0. Then choose whatever portfolio you like. There are quite a few cycles where the portfolio stayed flat for near 20 years.

 

[audreyh1]

And, since usually those dips don't last very long (stocks >generally< recover within about 3 years IIRC, esp with reinvested dividends), it wouldn't take a huge pile money to ease through a rough patch. And, if a big expense comes up (new roof, emergency car purchase, etc) when stocks are down, then having a slug of cash would be useful.

Another option for those who don't like setting aside the cash like this would be to look at the results of using the Clyatt 95% rule (which FIRECalc will model easily). It avoids year-to-year withdrawal changes of more than -5% (plus inflation), has a negligible impact on portfolio survival, and avoids the cash "drag."

Taking a closer look.......

If you look at the toughest runs for %remaining portfolio - like 1966 - you see the real portfolio stays down for well over a decade, eventually dropping to 54% of the initial value. This is a run using 3.25% withdrawal rate, 50% total stock market, 50% treasuries.

[Year	%Initial Portfolio Real
1966	92.6%
1967	95.2%
1968	95.2%
1969	85.8%
1970	84.5%
1971	87.1%
1972	90.6%
1973	76.2%
1974	61.5%
1975	68.5%
1976	68.6%
1977	61.3%
1978	60.7%
1979	57.9%
1980	58.5%
1981	53.9%
1982	61.0%
1983	65.2%
1984	67.1%
1985	73.4%
1986	83.2%
1987	79.5%
1988	82.8%
1989	87.7%
1990	83.1%
1991	93.7%
1992	93.7%
1993	95.6%
1994	93.2%
1995	105.5%

 

[audreyh1]

I have thought about this some more, and rediscovered what I saw before (and kept forgetting despite my self-proclaimed "superior" memory). That is there are periods where there was just no gain to be had no matter where you look. Cash or CD did not save you, because, well everything was like cash: all showed zero return!

Using data from the Federal Reserve in St. Louis, I found the following for the period of 1/1966 to 1/1983, a period of 17 years.

$1 invested in stocks would become $3.04.
$1 invested in 3-month T-bills would become $3.175.
$1 invested in 10-year constant maturity T-bond would become $2.34.

However, the cumulative inflation factor over this 17-year period is 3.08.

So, your initial $1 would look like this after 17 years with inflation adjustment:

Stock: $0.99
T-Bill: $1.03
T-Bond: $0.76

There's no point in keeping a lot of cash (CD)! Everything looks like CD (or slightly worse), when averaged out over this 17-year period.

Now, that makes everything easy. If your WR is 3%, then after 17 years you will have spent about 1/2 of your stash, and have 1/2 of it left. If your WR is 4%, then you have about 1/3 left. It would be less if you hold long bonds. Don't hold long bonds!

It's not the end of the world. Your portfolio probably will survive for another 13 years to make it a 30-year retirement. However, you do not have to worry about dying rich.

But cash did not give you a negative return when other things did. That is the point. I claim that it did save you, so to speak, because you had something you could live off of and did maintain itself in real terms when everything else was underperforming!

Cash is not always the worst investment or "drag". There are historical periods during which "Cash was King" as some financiers like to phrase it.

 

[samclem]

Taking a closer look.......

If you look at the toughest runs for %remaining portfolio - like 1966 - you see the real portfolio stays down for well over a decade, eventually dropping to 54% of the initial value. This is a run using 3.25% withdrawal rate, 50% total stock market, 50% treasuries.

Yep, that would be a tough run. While stocks generally return to their previous values within 3 years, it can be a lot longer if we are taking withdrawals. And some downturns are far worse. No reasonable cash buffer or clever withdrawal strategy will change the fact that a lot of sacrifices would be needed in some cases.
I try to imagine what my thoughts/state of mind would be while actually riding a line like that downward--with no true guarantee that things would pop back up before my portfolio went very low. One thing is for sure--if I had initially planned to take withdrawals of x% and adjust it annually for inflation, I'd quickly abandon that idea. Who would just blindly spend their portfolio to zero based on the hope that things couldn't get worse than they had in the past? Knowing that, it seems logical (to me, in muy situation) to just start out using use some variant of the "% of annual balance" method.

 

[NW-Bound]

But cash did not give you a negative return when other things did. That is the point. I claim that it did save you, so to speak, because you had something you could live off of and did maintain itself in real terms when everything else was underperforming!

Cash is not always the worst investment or "drag". There are historical periods during which "Cash was King" as some financiers like to phrase it.

Are you talking to me, someone who never has less than 20% in cash?

Some cash is always good. But looking at the period of 1966-1983, it was clear that the amount of cash I usually keep is way too much, even if it is in I-bonds. About 2-year worth of expenses, at most 3, should be enough.

Now, if someone has those I-bonds with 3%+inflation, I can see holding them nearly forever. Who needs stinkin' regular bonds with that return? But I have only the 1%+inflation bonds, and am now thinking I can do better with something else.

Cash is useful to deploy during flash crash like Brexit, or short V-shape recession like in 2009. But there are other ways one can "go in" more when such opportunities present themselves. The amount of cash I have is just way too much.

 

[NW-Bound]

Taking a closer look.......

If you look at the toughest runs for %remaining portfolio - like 1966 - you see the real portfolio stays down for well over a decade, eventually dropping to 54% of the initial value. This is a run using 3.25% withdrawal rate, 50% total stock market, 50% treasuries...

The portfolio did not stay down. It was just flat (with WR=0), no matter what your AA is, whether 100% stock or 100% bond. There's just no return, as I observed in the posts above.

Given that there's no investment gain anywhere, you spend 1/2 of your stash in 17 years at 3% WR (3% x 17 = 51%). One does not even need FIRECalc.

 

[daylatedollarshort]

I have thought about this some more, and rediscovered what I saw before (and kept forgetting despite my self-proclaimed "superior" memory). That is there are periods where there was just no gain to be had no matter where you look. Cash or CD did not save you, because, well everything was like cash: all showed zero return!

We're more into capital preservation than growth, so I look at the estimated real return from a TIPS ladder for my spreadsheet. Right now I'm using a .5% real return. The current rates are here:

Bloomberg Treasury Yields
https://www.bloomberg.com/markets/rates-bonds/government-bonds/us

If we make more than that it will be party time. At zero real returns for 30 years we can withdraw 3.33% per year (100 / 30 years = 3.33%). If we can eek out another .5% above a 0 real return then it is closer to 4 percent. But we will always keep out withdrawal rates under that because we still like to live below our means and will plan for living longer than that.

 

[audreyh1]

The portfolio did not stay down. It was just flat (with WR=0), no matter what your AA is, whether 100% stock or 100% bond. There's just no return, as I observed in the posts above.

Given that there's no investment gain anywhere, you spend 1/2 of your stash in 17 years at 3% WR (3% x 17 = 51%). One does not even need FIRECalc.

What is the point of looking at performance with no withdrawals? We are looking at historical real portfolio performance while withdrawing.

The one I ran the portfolio did go down in real terms. It got down to 54% of the original after 16 years - because you were drawing from it at the same time it was declining due to performance. A portfolio does not stay flat during years of 0% real market returns, because your withdrawals make it shrink. Your spending would have to shrink as well assuming your spending power was linked to the same inflation, which I think is a reasonable assumption. So after a gradually declining portfolio, you still go through a long period of spending 61% to 54% of your initial spending - you had better have plenty of discretionary spending you can just drop and do without. If you don't want to draw the portfolio down even farther, that is.

And just because a portfolio was flat over a given period, doesn't mean that the real performance wasn't negative over part of that time. That's why you can't just divide an initial portfolio by a withdrawal rate to decide how long to get down to some number.

 

[NW-Bound]

Here's a simpler way to look at it. From my earlier post,

Using data from the Federal Reserve in St. Louis, I found the following for the period of 1/1966 to 1/1983, a period of 17 years.

$1 invested in stocks would become $3.04.
$1 invested in 3-month T-bills would become $3.175.
$1 invested in 10-year constant maturity T-bond would become $2.34.

However, the cumulative inflation factor over this 17-year period is 3.08.

So, your initial $1 would look like this after 17 years with inflation adjustment:

Stock: $0.99
T-Bill: $1.03
T-Bond: $0.76
...

This means that the portfolio (without withdrawal) is exactly flat once inflation is accounted for. Your spending should be also adjusted for inflation. So, let's think about the dollar in 1966 valuation terms.

You have $1M to start out with. It stays constant (in 1966 terms).

You keep spending 3% a year, again in 1966 terms. So, in 10 years, you spend 30%, and so forth. In 17 years, you spend 3 x 17 = 51%. No matter what the nominal value is, your portfolio has shrunken to 1/2 of its original purchasing power.

What I found as very interesting is that it was the same whether you do 100% stock or 100% T-Bill in that 17-year period.

You can keep some "off-portfolio" money, and if it also has no gain, your total on+off portfolio is still the same. It can be 100% off-portfolio or 100% on-portfolio, your total amount still goes down to 1/2 in 17 years.

Now, one can look closer and see if an opportunist could rebalance between the two stock and T-bill, as they probably were not always in sync. He could pick up a bit of gain there. This is something I am going to look at for myself next, using the raw data from the Fed.