Retiring In Secular Cycles

[wabmester]

This came up on raddr's board last May:
http://www.raddr-pages.com/forums/viewtopic.php?t=2546

Wow, that's great work, bpp. Let's hope it doesn't happen here. Lots of good insights on that thread.

BTW, Swedroe had some data last year on the recent ScV outperformance, and I believe it was the highest in history (or at least since WWII). Scary all around. And I do believe that F&F didn't find much of a small-cap premium internationally.

Yes, I assumed 'yield' for the quick analysis since the presumption would be that today's investor would be "stuck" with current yields. Even if there were total return gain/losses, the reinvestment rate would be at market yields and thus would imbed the now current yield.

That's a good point -- bond returns are more predictable than stock returns. It seems a little odd to mix current bond yield with historic stock returns, but I like the insight. Maybe you'd get a different SWR result if you also assumed the current embedded inflation estimate in those bonds (assuming you used historic CPI data in your analysis).

 

[crestmont]

While not showing P/Es, this chart from FPA 1 IMHO shows pretty well that the 10+% returns averaged since ~1990 are atypical, and if past is prologue, we should look for ~6.7% to return.

Note that the chart is Total Return (Index + Dividends) adjusted for inflation. Since there are only 3 sources of return from stocks: (1) EPS Growth + (2) Dividend Yield + (3) P/E Change, the chart is actually showing us Real EPS Growth + Dividends (since the impact of P/E changes, if any, over 135 years would be minimal). Given that Real EPS growth tracks economic growth over time, and since Real EPS grew at just over 3% over that period, we have the first part of the total. Dividend yields averaged close to 3.5%. Thus the the total is 6.7%, in line with the chart. The key thing to note about this is that the chart is reflecting rational economics, not random finance. Going forward, from today's level of valuation and expected economic growth, the outlook is closer to 6.5% or a little less.

That's a good point -- bond returns are more predictable than stock returns. It seems a little odd to mix current bond yield with historic stock returns, but I like the insight. Maybe you'd get a different SWR result if you also assumed the current embedded inflation estimate in those bonds (assuming you used historic CPI data in your analysis).

Yes, using today's yield was an assumption since the retiree would be buying long-term bonds in the current interest rate environment and so any scenario with higher or lower yields would not be applicable. Also, since the anaysis really focuses on the risks in the top quartile of P/Es--periods when inflation was very low--bond yields in those periods would have been close to 5%. Had the bond yields varied with inflation changes, I think the impact would reflect even better success (not a lower risk of failure) in lower quartile periods since the bonds would be acquired at higher yields. Thanks for your comment.

 

[FIRE'd@51]

Yes, using today's yield was an assumption since the retiree would be buying long-term bonds in the current interest rate environment and so any scenario with higher or lower yields would not be applicable. Also, since the anaysis really focuses on the risks in the top quartile of P/Es--periods when inflation was very low--bond yields in those periods would have been close to 5%. Had the bond yields varied with inflation changes, I think the impact would reflect even better success (not a lower risk of failure) in lower quartile periods since the bonds would be acquired at higher yields.

I think this clearly reinforces the notion of having a much shorter duration in the fixed income portion of one's portfolio. Using laddering-techniques, inflation-protected securities, and money-market funds allows the yield earned on the fixed income portion to respond more quickly to changes in inflation. What really killed portfolios in the 1970's and early 80's was higher inflation coupled with having a whole lot of long-term bonds yielding less than five percent. Perhaps that is the bigger risk to a portfolio today. The Ibbotson data shows that, from 1926-1982, the real return on T-bills was about zero. Since then it's been about 2%. Today we have low nominal interest rates, especially at the intemediate and long end of the yield curve. However, real rates are pretty high in historic terms. Even TIPS allow you to lock-in real rates of 2.3-2.4%. So I would argue that, properly managing the fixed income portion of one's portfolio should offset much of the risk of somewhat higher-than-average equity valuations. As it was in the 60's, the major risk to today's retiree is a big increase in inflation. Historically low inflation is what allowed us to weather the big equity decline of 2000-2002.

 

[justin]

Please keep in mind that the most-often quoted data--Ibottson's Annual Yearbook--only reflects long-term returns of 10% because the series starts in 1926 when P/Es were 10.2 and ends when P/Es are almost twice as high. Starting P/E impacts the dividend yield component of total return as well as the capital gain component. Had P/Es at the start been the same as today, long-term returns would have been (necessarily) less than 7%. That alone would be a major change to conventional thinking. The starting point has a major impact.

Wait a sec... I may be doing the math wrong, but the expansion of the PE ratio from 10 to ~18 over the last 81 years seems to account for only 0.8% of annual return. The other 9.2% came from "something else". EPS growth and divs.

Based on 10% returns for 81 years, the 1926 values are up 225224%. If the PE would have remained at 10, then the values would be up 125124%.

My point is that over really long periods of time, the relevancy of PE ratios slowly fades away.

One additional point - let's say I'm one year from FIRE, and I have 95% of my target portfolio value in the bank. PE's are currently at 18. 1 year from now, PE's have gone to 36, and earnings growth was zero. My portfolio just doubled in value, so I can afford to take 1/2 of my originally planned "4% SWR" and still get by. So now what is the survivability of a 2% SWR with starting PE's of 36? That's the question I would be interested in finding out. My portfolio value will increase significantly with an expansion of the PE ratio (all else being equal). This will affect what % I'll need to take to live on.

 

[yakers]

1 year from now, PE's have gone to 36, and earnings growth was zero. My portfolio just doubled in value, so I can afford to take 1/2 of my originally planned "4% SWR" and still get by. So now what is the survivability of a 2% SWR with starting PE's of 36? That's the question I would be interested in finding out. My portfolio value will increase significantly with an expansion of the PE ratio (all else being equal). This will affect what % I'll need to take to live on.

Hey, that’s a good point, maybe I should borrow the copyrighted Ding, Ding, Ding exclamation . I usually don’t look at it the way you suggested because I am not going for a target amount in my portfolio. I am looking at my age and anticipated pension and the rest of my portfolio I just look at as having ‘this much’ on my retirement date and 4% should be a SWR. But what you said makes a lot of sense. If I am looking at a portfolio of $400k and a PE <20 and the PE goes way over 20 but the portfolio goes way over $400k, what is the problem? I have not seen it stated this way. I think Ho%us would find that difficult to deal with.

 

[justin]

If I am looking at a portfolio of $400k and a PE <20 and the PE goes way over 20 but the portfolio goes way over $400k, what is the problem?

That's the way I figure it. I'm shooting for ~$1.5 million. That will give me $60k/yr based on a 4% SWR. That's all I need to fund my FIRE plans.

Well, if the PE's double in my last couple years of investing (and earnings growth remain relatively flat), then the portfolio goes from $1.5 million to $3 million. I still only need $60k/yr for my expenses. That equates to a 2% SWR on the $3 million portfolio.

I don't know how to quantify the effect of PE on SWR, or how reliable or accurate any quantification would be. I feel pretty good about the numbers FIREcalc produces (ie - FIREcalc is telling you if you would have survived historically). But I also keep an eye on PE ratios for a general valuation measure. I think they are pretty useful when they are at extremes, and not very useful when they are in the "middle" territory.

 

[unclemick]

"Dividends Don't Lie"
Geraldine Weiss Investment Quality Trends

And no - Yogi Berra didn't really say 'dividends are the same as real money'

But he came close!

heh heh heh heh heh heh heh heh - slept in to honor the snow outside.

 

[wabmester]

Yes, using today's yield was an assumption since the retiree would be buying long-term bonds in the current interest rate environment and so any scenario with higher or lower yields would not be applicable. Also, since the anaysis really focuses on the risks in the top quartile of P/Es--periods when inflation was very low--bond yields in those periods would have been close to 5%. Had the bond yields varied with inflation changes, I think the impact would reflect even better success (not a lower risk of failure) in lower quartile periods since the bonds would be acquired at higher yields.

That's a reasonable statement. Much more reasonable than:

The results are that the current level of bond yields (5% or so) reduces the success rate (assuming annual rebalancing) at all stock/bond mix ratios.

Which might scare people away from bonds (or at least reinforce the gung-ho stock allocators).

Bottom-line: long-term bonds probably aren't a safe haven during times of high stock valuations. Maybe cash is king?

I wonder if there have been other periods in history with a global liquidity glut making almost all asset classes "expensive" on a global scale. What are the safe havens during times like we have today?

 

[haha]

I wonder if there have been other periods in history with a global liquidity glut making almost all asset classes "expensive" on a global scale. What are the safe havens during times like we have today?

Given that US$ and other currencies are fiat, and all income producing assets are high priced, there really isn't an obvious "safe haven". Maybe exotic things for rich people, like ranches in the pampas. But for one who needs income, therefore can't make big bets that might take years and large new capital injections to carry out, looks like slim pickings. TIPS aren't bad, but real interest rates are not as high as they have been, even in the recent past. So there is an opportunity for capital depreciation there, though not absolute loss.

It's a time for speculators, in that we are essentially forced to speculate by the poor passive returns to capital being offered in the marketplaces.

Justin’s argument above about PEs going from 18 to 36, therefore everything still copasetic on the SWR front is a speculator’s argument. “Hell boy, you think 18 is high, just watch!”

And why not? Why not go to 100; surely someone created the intellectual framework for that during the last big boom.

Ha

 

[yakers]

Justin’s argument above about PEs going from 18 to 36, therefore everything still copasetic on the SWR front is a speculator’s argument. “Hell boy, you think 18 is high, just watch!”


Ha

The element Justin added was that he would use a lower SWR because his portfolio has increaded significantly. I found that a interesting idea.

 

[Nords]

What are the safe havens during times like we have today?

I bonds?

Maybe they're just a "safer" haven. Like bungee jumping with a parachute.

 

[haha]

That is *****' idea in a nutshell, and of course absolutely contra to the Religion of Retirement Calculators. None of it really makes any sense when you try to push it beyond it's obvious useful purpose- to help people (including many FPs) realize that one can't just take total smoothed annual rturn, subtract inflation and what is left over is draw. Volatility kills that plan.

That is really all there is here.

Ha

 

[justin]

Justin’s argument above about PEs going from 18 to 36, therefore everything still copasetic on the SWR front is a speculator’s argument. “Hell boy, you think 18 is high, just watch!”

And why not? Why not go to 100; surely someone created the intellectual framework for that during the last big boom.

If PE's went to 36 today, I would probably make a tactical asset allocation decision to reduce my new contributions and/or existing holdings in equities. At PE = 18, I feel ok as a long term investor (75+ years maybe). I'm certainly not going to get rich with 5.x% bond yields!

The tricky thing about using PE as an indicator is that the E could get cut in half (which may happen if we hit a recession). Does the PE ratio get cut in half? Go down 60%? Or does the P drop 50%, thereby maintaining the same PE of 18? My guess is that when the E starts dropping rapidly, the PE will drop too, thereby amplifying the drop in P. Good for me as someone in the accumulation phase, not so much for those FIRE'd in the withdrawal phase.

 

[justin]

The element Justin added was that he would use a lower SWR because his portfolio has increaded significantly. I found that a interesting idea.

By my thinking, I would not HAVE to use a lower SWR necessarily, but my actual WR might be lower if I have a sudden unexpected last minute boost to my portfolio value.

I would qualitatively consider it more risky to retire after a 3-4 year very strong bull market where I experienced 20%+ returns each year AND I was basing my SWR on my newly inflated portfolio value.

 

[Animorph]

Which is all Crestmont is saying as well. It's riskier to retire just after a bull market (high P/E's) than just after a bear market (low P/E's).

Dan

 

[terminator]

My guess is that when the E starts dropping rapidly, the PE will drop too, thereby amplifying the drop in P.

My own qualitative observations while doing research tells me it depends on the strength of the company (size, debt, market share/position). Large blue-chips seem to tend to keep the same P/E as earnings go higher (and quite often the P doesn't change much so the P/E goes higher) because investors "know" that the earnings will come back up eventually and the company is paying dividends in the meantime. Smaller companies are more likely to have their P/E's shrink since their earnings are more volatile overall and they are less likely to be in as strong a position during a downturn as the large blue-chips. But then again, after a downturn shakes out then there would be more "value" stocks in the smaller ranks. Just my observations.

 

[youbet]

Which is all Crestmont is saying as well. It's riskier to retire just after a bull market (high P/E's) than just after a bear market (low P/E's).

Dan

Not exactly....... Say in 2003 I had $1,000,000 and planned to retire in 2007 with an estimated $1,300,000. But, to my pleasant surpirse, a nifty bull market takes my portfolio value to $1,500,000 or $200,000 more than I estimated. I go ahead and retire in 2007. I don't feel any more at risk than if I had retired with the $1,300,000 that would have resulted from a less positive market.

 

[Animorph]

Youbet,

But you are taking more risk, if you believe Crestmont. FIRECalc may say that you have a 95% chance of survival on average, but Crestmont notes that all of those 5% failures have come when starting retirement at high P/E's. If you consider only those retirement start years with similar high P/E's, your success rate looked more like 80%. Even if you use bonds, your stock portion is still subject to those lower returns. This is one simple way of trying to predict future returns, unless you think the P/E is going to be 40 when you die.

I was happy it wasn't worse, and we are on the line between the quartiles, but I don't expect to see much P/E gain during retirement.

Dan

 

[youbet]

Dan.....

Perhaps I need to explain further. The hypothetical situation I gave is similar to what actually happened to me. In 2003, I felt we were FI and I was considering retiring. Being somewhat financially conservative ( chickensh!t) by nature, I dragged my feet and waited until June, 2006, when my company helped me make the decision. The market did better, by far, between 2003 and mid-2006 than I estimated it would. My portfolio was significantly larger than I planned it would be. I asked myself "since I have considerably more money than I planned on having at this point, should I retire anyway?" And I answered myself "YES!"

I don't withdraw any more in absolute terms than I would have if my portfolio had grown more slowly. All my plans were in dollars, not percentages. So, despite the higher PE at retirement, I feel I'm taking no additional risk as I have more money and my WR is lower as a percentage.

 

[New Thinking]

If you were to view equities at this "point in time", shouldn't you also view the bond component as a point in time as Crestmont suggests? You are not going to get price appreciation over the next X years..Look at Business Week's cover article this week, "It's a Low,Low,Low, Low-Rate World".

And although human behavior could cause one to lower a SWR, human behavior will cause more folks to buy and sell at the wrong times and do more damage..Not the profile of many of you, but the average individual..Look at the Dalbar site for their annual mutual fund returns of the average investor over the last 20 years.

Dalbar shows that the following returns over the twenty-year period of 1986-2005:

· S & P 500 return of 11.9%
· Average Equity Fund Investor earned 3.9%
· Long-Term Shearson Government Bond Index 9.7%
· Average Fixed Income Investor earned 1.8%7


This is the real danger for the masses of future retirees. Its not all investment expenses but fear and greed driving behavior.

 

[Tadpole]

I am enjoying this.

Should the SWR be set at a level commensurate with the value of the portfolio at the end of the last bear market adjusted to the value the portfolio would have had if it had been invested entirely in bonds during the bull cycle?

 

[Gone4Good]

This came up on raddr's board last May:
http://www.raddr-pages.com/forums/viewtopic.php?t=2546

Maybe I'll do an update through 2006 at some point, but the basic results won't change: 3% and diversification starting in 1990 would have left one intact (so far).

You'll notice that he also has a 50/50 stock/bond mix. In the first pass, without international diversification, and a 60/40 split the 3% withdrawal scenario "survived" but in year 15 you were pulling 10% withdrawals - even after some pretty good equity returns over the last few years.

I put together my own numbers using the Nikkei this afternoon (before seeing this link) and my calculations tie back to the results posted therein. Some interesting notes from my calculations:

1) Higher bond allocations were always better.
2) Retiring 3 years earlier in 1987 (allowing for a 100% increase in equity prices between 1987 & 1990) helped, but not as much as you might think. And higher bond allocations were still always better.
3) A retiree in Jan 1990, using a 50/50 mix, 4% initial withdrawal rate, and a flexible withdrawal would have to reduce his real standard of living by 45% in order to keep a 4% withdrawal rate by 2006.
4) In the same scenario as #3, an 80/20 portfolio would require a 73% reduction in standard of living
5) A 3% withdrawal rate, 50/50 mix, requires a 23% reduction in standard of living to have a 4% withdrawal rate in '06.
6) A 3% withdrawal rate, 80/20 mix, requires a 58% reduction in living standards.

"Safe Withdrawal Rates" defined as no reduction in real spending and a 4% withdrawal rate in 2006 is as following for the following equity/debt mix:

50/50 2.4%
60/40 2.1%
80/20 1.4%
100/0 0.94%

It does make you take a second look at the risk/reward trade off between the security of holding more bonds versus the foregone income potential of higher equity ownership.

It also makes you think twice about FIRECalc as a "worst case" simulator.

 

[DRiP Guy]

It also makes you think twice about FIRECalc as a "worst case" simulator.

Makes me think twice about investing in the Nikkei !


(j/k -- obviously today's best wisdom seems to be to have a fair amount of international exposure)

 

[wabmester]

Makes me think twice about investing in the Nikkei !

Japan is coming off a 15-year-long "recession." They've seen deflation, which we haven't seen in 60+ years. And their GDP is growing faster than ours. I think Japanese who ER now will probably do OK. Maybe it's time to think about going 100% NIKKEI.

 

[Gone4Good]

Japan is coming off a 15-year-long "recession." They've seen deflation, which we haven't seen in 60+ years. And their GDP is growing faster than ours. I think Japanese who ER now will probably do OK. Maybe it's time to think about going 100% NIKKEI.

Still a 38 P/E.