Indicators and Retirement Portfolios

Research on Safe Withdrawal Rates

Moderator: hocus2004

Post Reply
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Indicators and Retirement Portfolios

Post by JWR1945 »

Indicators and Retirement Portfolios

I am very interested people's thoughts regarding indicators and their applications related to retirement.

So far, we have found that P/E10 gives us a very useful tool. Tobin's q is almost as good. Dividend yields are critically important. P/E, as opposed to P/E10, has not looked too good. After I read a recent article, it struck me that P/E might be exceedingly valuable, especially for real retirees with real portfolios and also for those close to retirement. P/E does a great job of predicting the downside risk over the next ten years. It does not predict the upside very well, but the downside risk is critically important to retirees.

Refer to this post:
http://nofeeboards.com/boards/viewtopic ... 695#p11695
and its link to the research paper:
http://www.fpanet.org/journal/articles/ ... -art10.cfm

Right now, we know that we can use long-term TIPS to help us coast through bad times in the stock market. But who really wants to be on the sidelines for 10 or 15 years without buying anything? It may turn out that someone's magic number (e.g., a P/E10 threshold of 12) won't be reached until longer. It may be that occasions will arise when the short-term P/E (based on a single year of trailing earnings) signals an opportunity even though P/E10 does not. To the extent that the short-term P/E correctly indicates that the downside risk is low, a real life retiree could buy some shares of stock with a reasonable degree of safety so as to avoid being locked out of the market altogether.

It strikes me that each indicator has its own strengths that can help to compensate for the weaknesses of others. It may be that a particular indicator is especially good over a specific period of time. For example, P/E10 is better than Tobin's q when estimating Safe Withdrawal Rates for the next thirty years. It may well be that Tobin's q gives us better information in the earlier years of retirement. Since the first decade or so (i.e., eleven years, historically) is the most important in terms of portfolio success in retirement, it might turn out that Tobin's q gives us invaluable supplemental information. In terms of how this could plausibly come about, it might be that Tobin's q accurately indicates the safety of dividends. I do not know whether it does or does not. But I do know that a cut in dividends in the first two or three years causes us to revise our Safe Withdrawal Rate calculations downward.

My emphasis on this board is in terms of retirement and the distribution stage. Undoubtedly, the usefulness of this kind of information extends into the pre-retirement years, especially those very close to retirement. You may be willing to accept a large downward fluctuation in your portfolio balance ten or fifteen years before retirement, but you do not want to do so later.

Have fun.

John R.
hocus
Moderator
Posts: 435
Joined: Mon Dec 02, 2002 12:56 am

Post by hocus »

Who really wants to be on the sidelines for 10 or 15 years without buying anything?

I presume that by the phrase "on the sidelines," you mean to be out of stocks. Personally, I would not advise hardly anyone to be out of stocks for that long a time-period. Personally, I would not advise hardly anyone to be entirely out of stocks for even a much shorter time-period.

My hope is that our work will help to restore some balance to investing discussions. When things have reached a point where there are people putting out studies claiming that a 74 percent allocation to S&P stocks is "optimal" for all possible investors at all possible times and are not being laughed out of the house for doing so, the average investor's perceptions have gotten seriously out of whack in a tulip buib mania sort of way. One concern I have about the extreme views re stock investing that have been expressed in the past 10 years is that they may cause a powerful swing in the opposite direction in coming years when people find that "research" of the conventional SWR study variety (but not just that) was build on sand. When people find out how they have been tricked by a lot of people who have put forward all sorts of views that "everyone knows are true" (but which are at great variance to what a reasoned study of the historical data reveals), I think we may see middle-class investing in stocks drop to low levels.

I hope that the work we do at this board will show that it is as much a bad idea to take one's stock allocation too far on the low side in a bear market as it is to take it too far on the high side in a bull market. There's no question but that the idea that a 74 percent S&P stock allocation is always safe is loony tunes. But I worry that the same emotions that caused people to buy into that line at the time when stocks were booming will cause people to buy into the line of getting out of stocks altogether when the bear market's version of intercst comes along to tell us that the historical data tells us that no middle-class investor can ever afford to be investing in any stocks whatsoever

I don't mean to ascribe these extreme views to you, JWR1945. I know that you do not hold them. But I expect that one thing we are going to need to do in working through recommended implementations of the insights that have been developed at this board is to provide strategies for how to take what the historical data reveals and apply it in balanced and moderate ways. I tend to think that even at times at which the SWR for stocks is exceedingly low, most middle-class investors should have at least a small allocation to stocks, if only to keep their heads in the game while prices find their way to more appealing levels.

Undoubtedly, the usefulness of this kind of information extends into the pre-retirement years, especially those very close to retirement.

I think that's right. In my view, the usefulness extends not only to those retired or close to being retired. I think that it is too narrow a view of the SWR analytical tool to think of it as providing value only to those retired or about to become retired (although I presume those were the groups that were foremost in the minds of the people who took the early stabs at development of the tool)

To the extent that SWR analysis reveals something about the intrinsic value of a stock purchase, it provides an insight that is valuable not only to the soon-to-be-retiree, but to all others investors as well. The tool does not provide perfect and complete information on intrinsic value, of course. But I think it provides meaningful insights re this question. All of the various types of stock investors should want to know the SWR that applies for stocks at the times at which they are planning stock purchases, in my view.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Who really wants to be on the sidelines for 10 or 15 years without buying anything?

I presume that by the phrase "on the sidelines," you mean to be out of stocks. Personally, I would not advise hardly anyone to be out of stocks for that long a time-period. Personally, I would not advise hardly anyone to be entirely out of stocks for even a much shorter time-period.
I agree. In fact, we have emphasized increasing the number of options available to the investor so that he can do what is best for himself and his own personal circumstances.

One of the first things that I tell people is that they have time. They do not have to invest right away. They have time to learn. They can invest in a way that makes them comfortable. They do not have to be rushed into anything.

We have seen that a 100% TIPS portfolio could have gotten you through the worst investment period, retirements beginning in 1959-1973, before buying stocks. You would have still had enough of your principal left to take advantage of the opportunities that followed. That is an extreme case, but it supports the idea that you do not have to rush. You can wait.

My efforts are to expand one's choices and to reduce critical sensitivities. It is worth remembering that there is always a risk associated with any analysis, no matter how well done. We need to be able to handle the results of such a failure. Introducing more indicators, when done correctly, should help us make the right changes if our original choices prove incorrect.

As an example of such a failure, I refer to the Retire Early results when the analysis was expanded to include starting in months other than January. The optimal stock allocation percentages jumped all over the place. Ignore the other problems with the study for purposes of this discussion: that was a biggie. IIRC, the correct allocation changed from 74% (or so) to 60% (or was it 50%?). It was a biggie and it was something that retirees should have been warned about.

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Two new indicators

Hardware Database Rate Tables and Summary


I have listed two sets of tables. One is ordered according to the Hardware Database Rates (HDBR) and the other is ordered in accordance with the lower of P/E and P/E10.

These are the same Hardware Database Rates from 1921-1980 that I have listed before. They were determined from the FIRECalc calculator in increments of 0.2%. All portfolios survived at least thirty years at the rates shown. There was at least one portfolio that failed within thirty years when the rate was increased by 0.2%. The portfolio consisted of 80% stocks (i.e., the S&P 500) and 20% commercial paper. The expense ratio was 0.20%. Withdrawals were changed each year to match inflation. The CPI was used for inflation adjustments.

I have listed the lower of P/E and P/E10 and their ratio ( [P/E10] / [P/E] ). All terms are real values, adjusted for inflation. The price is the most recent value of the S&P 500 index (after adjusting for inflation), E10 is the average of the previous ten years of earnings (trailing, reported earnings) and E is the earnings for the current year (trailing, reported earnings). Because P and E both occur in the same year, their ratio P/E is the same whether or not they are adjusted for inflation. Values of P (both nominal and real), E (both nominal and real) and P/E10 are listed in Yale Professor Shiller's database.

The ratio ( [P/E10] / [P/E] ) equals ( E / E10 ), where both E and E10 are adjusted for inflation. Historically, real earnings have grown slightly more than 1% per year (when averaged over long time periods). Because E is from the current year and E10 is from the past ten years, the ratio ( E / E10 ) should average close to 1.05.

Observations drawn from the HDBR ordering.

Look at the ratios. The table with the highest HDBRs and the middle HDBRs both have ( [P/E10] / [P/E] ) ratios greater than 1.00 twelve times (and less than 1.00 eight times). The table with the lowest HDBRs have ratios greater than 1.00 fourteen times (and less than 1.00 six times). This is not dramatic enough to form a strong conclusion.

Now look at the last nine entries on the third table. The ratio was greater than 1.00 in every single case that the Hardware Database Rate was 4.6% or lower. The P/E ratio was falling relative to P/E10. Earnings were increasing relative to earnings in the previous decade.

I do not consider this sufficient to draw a strong conclusion that would support a prediction. I do consider this finding sufficient to reject an alternative assertion. That is, just because the price to earnings ratio is falling (P/E, relative to P/E10) or because earnings are rising (E, relative to E10), you cannot assume that it is safe to retire. The price to earnings ratio was falling during the most stressful conditions seen in the historical database.

Observations from the alternative ordering.

Look at those tables that I ordered according to the lower of P/E and P/E10. For those twenty years best valuations, the Hardware Database Rates varied between 7.0% and 11.6%. For those in the middle range, the Hardware Database Rates varied between 4.4% and 9.4%. For those with the highest valuations (i.e., highest prices relative to E or E10), the rates varied between 4.0% and 6.8%.

The comparable ranges based on P/E10 alone were 7.2% to 11.6%, 5.6% to 9.4% and 4.0% to 5.8%.

We see that P/E10 alone did better than using the lower of P/E and P/E10.

Careful examination of the tables at the best valuations shows that there might be a benefit to using the lower of P/E and P/E10 when P/E10 is very low. Otherwise, I have not discerned an advantage to this method of ordering.

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

These are ordered by HDBR (80% stocks / 20% commercial paper).

Code: Select all

Year HDBR(80%) P/E10  P/E  Dividends  lower  ratio
1921   11.6%    5.1    9.4   7.11%     5.1   0.54
1949   11.0%   10.2    6.6   6.18%     6.6   1.55
1948   10.8%   10.4    9.0   5.69%     9.0   1.16
1980   10.4%    8.8    7.4   5.18%     7.4   1.19
1922   10.2%    6.2   22.6   6.35%     6.2   0.27
1950   10.2%   10.7    7.2   6.84%     7.2   1.49
1979    9.6%    9.2    7.9   5.12%     7.9   1.16
1924    9.4%    8.0    9.0   6.02%     8.0   0.89
1947    9.4%   11.4   13.5   4.69%    11.4   0.84
1951    9.4%   11.8    7.5   6.98%     7.5   1.57
1943    9.2%   10.1    9.7   5.90%     9.7   1.04
1978    9.2%    9.2    8.3   5.22%     8.3   1.11
1923    9.0%    8.1   12.5   5.74%     8.1   0.65
1942    9.0%   10.1    8.0   7.87%     8.0   1.26
1952    9.0%   12.5   10.0   5.86%    10.0   1.25
1954    9.0%   12.0   10.1   5.73%    10.1   1.19
1944    8.6%   11.0   12.7   5.19%    11.0   0.87
1953    8.6%   13.0   10.9   5.40%    10.9   1.19
1925    8.4%    9.6   11.0   5.26%     9.6   0.87
1933    8.4%    8.7   17.2   6.98%     8.7   0.51

Code: Select all

Year HDBR(80%) P/E10  P/E  Dividends  lower  ratio
1945   8.2%   11.9   14.4   4.77%    11.9   0.83
1975   8.2%    8.9    8.3   4.99%     8.3   1.07
1932   8.0%    9.3   14.0   9.55%     9.3   0.66
1926   7.6%   11.3   10.1   4.82%    10.1   1.12
1935   7.6%   11.4   16.2   4.85%    11.4   0.70
1977   7.4%   11.4   10.4   3.97%    10.4   1.10
1927   7.2%   13.1   10.9   5.19%    10.9   1.20
1976   7.2%   11.1   11.8   3.80%    11.1   0.94
1941   7.0%   13.9   10.0   6.41%    10.0   1.39
1958   7.0%   13.7   12.5   4.33%    12.5   1.10
1946   6.8%   15.6   19.2   3.70%    15.6   0.81
1955   6.8%   15.9   12.6   4.34%    12.6   1.26
1971   6.8%   16.4   18.1   3.35%    16.4   0.91
1938   6.6%   13.5   10.5   7.02%    10.5   1.29
1934   6.2%   13.0   23.7   4.19%    13.0   0.55
1939   6.0%   15.5   18.8   4.10%    15.5   0.82
1940   6.0%   16.3   13.2   5.06%    13.2   1.23
1957   6.0%   16.7   13.3   3.82%    13.3   1.26
1928   5.8%   18.8   15.5   4.43%    15.5   1.21
1956   5.8%   18.2   12.1   3.78%    12.1   1.50

Code: Select all

Year HDBR(80%) P/E10  P/E  Dividends  lower  ratio
1974   5.8%   13.5   11.7   3.53%    11.7   1.15
1931   5.6%   16.7   17.0   6.07%    16.7   0.98
1936   5.4%   17.0   17.9   3.50%    17.0   0.95
1959   5.4%   17.9   18.8   3.15%    17.9   0.95
1960   5.2%   18.3   17.1   3.21%    17.1   1.07
1961   5.2%   18.4   18.6   3.26%    18.4   0.99
1963   5.0%   19.2   17.7   3.28%    17.7   1.08
1930   4.8%   22.3   13.9   4.47%    13.9   1.60
1962   4.8%   21.1   21.3   2.93%    21.1   0.99
1970   4.8%   17.0   15.8   3.50%    15.8   1.08
1972   4.8%   17.2   18.0   2.98%    17.2   0.96
1937   4.6%   21.6   16.8   4.17%    16.8   1.29
1964   4.6%   21.6   18.8   3.00%    18.8   1.15
1973   4.6%   18.7   18.1   2.67%    18.1   1.03
1967   4.4%   20.4   15.3   3.41%    15.3   1.33
1929   4.2%   27.0   17.8   3.46%    17.8   1.52
1965   4.2%   23.2   18.7   2.92%    18.7   1.24
1968   4.2%   21.6   17.7   3.08%    17.7   1.22
1969   4.2%   21.1   17.6   3.01%    17.6   1.20
1966   4.0%   24.0   17.8   2.93%    17.8   1.35

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

These are ordered by the lower of P/E10 and P/E (HDBR for 80% stocks / 20% commercial paper).

Code: Select all

Year HDBR(80%) P/E10  P/E  Dividends lower  ratio
1921   11.6%    5.1    9.4   7.11%    5.1   0.54
1922   10.2%    6.2   22.6   6.35%    6.2   0.27
1949   11.0%   10.2    6.6   6.18%    6.6   1.55
1950   10.2%   10.7    7.2   6.84%    7.2   1.49
1980   10.4%    8.8    7.4   5.18%    7.4   1.19
1951    9.4%   11.8    7.5   6.98%    7.5   1.57
1979    9.6%    9.2    7.9   5.12%    7.9   1.16
1924    9.4%    8.0    9.0   6.02%    8.0   0.89
1942    9.0%   10.1    8.0   7.87%    8.0   1.26
1923    9.0%    8.1   12.5   5.74%    8.1   0.65
1975    8.2%    8.9    8.3   4.99%    8.3   1.07
1978    9.2%    9.2    8.3   5.22%    8.3   1.11
1933    8.4%    8.7   17.2   6.98%    8.7   0.51
1948   10.8%   10.4    9.0   5.69%    9.0   1.16
1932    8.0%    9.3   14.0   9.55%    9.3   0.66
1925    8.4%    9.6   11.0   5.26%    9.6   0.87
1943    9.2%   10.1    9.7   5.90%    9.7   1.04
1952    9.0%   12.5   10.0   5.86%   10.0   1.25
1941    7.0%   13.9   10.0   6.41%   10.0   1.39
1954    9.0%   12.0   10.1   5.73%   10.1   1.19

Code: Select all

Year HDBR(80%) P/E10 P/E  Dividends  lower  ratio
1926   7.6%   11.3   10.1   4.82%   10.1   1.12
1977   7.4%   11.4   10.4   3.97%   10.4   1.10
1938   6.6%   13.5   10.5   7.02%   10.5   1.29
1953   8.6%   13.0   10.9   5.40%   10.9   1.19
1927   7.2%   13.1   10.9   5.19%   10.9   1.20
1944   8.6%   11.0   12.7   5.19%   11.0   0.87
1976   7.2%   11.1   11.8   3.80%   11.1   0.94
1947   9.4%   11.4   13.5   4.69%   11.4   0.84
1935   7.6%   11.4   16.2   4.85%   11.4   0.70
1974   5.8%   13.5   11.7   3.53%   11.7   1.15
1945   8.2%   11.9   14.4   4.77%   11.9   0.83
1956   5.8%   18.2   12.1   3.78%   12.1   1.50
1958   7.0%   13.7   12.5   4.33%   12.5   1.10
1955   6.8%   15.9   12.6   4.34%   12.6   1.26
1934   6.2%   13.0   23.7   4.19%   13.0   0.55
1940   6.0%   16.3   13.2   5.06%   13.2   1.23
1957   6.0%   16.7   13.3   3.82%   13.3   1.26
1930   4.8%   22.3   13.9   4.47%   13.9   1.60
1967   4.4%   20.4   15.3   3.41%   15.3   1.33
1928   5.8%   18.8   15.5   4.43%   15.5   1.21

Code: Select all

Year HDBR(80%) P/E10 P/E  Dividends  lower  ratio
1939   6.0%   15.5   18.8   4.10%   15.5   0.82
1946   6.8%   15.6   19.2   3.70%   15.6   0.81
1970   4.8%   17.0   15.8   3.50%   15.8   1.08
1971   6.8%   16.4   18.1   3.35%   16.4   0.91
1931   5.6%   16.7   17.0   6.07%   16.7   0.98
1937   4.6%   21.6   16.8   4.17%   16.8   1.29
1936   5.4%   17.0   17.9   3.50%   17.0   0.95
1960   5.2%   18.3   17.1   3.21%   17.1   1.07
1972   4.8%   17.2   18.0   2.98%   17.2   0.96
1969   4.2%   21.1   17.6   3.01%   17.6   1.20
1963   5.0%   19.2   17.7   3.28%   17.7   1.08
1968   4.2%   21.6   17.7   3.08%   17.7   1.22
1929   4.2%   27.0   17.8   3.46%   17.8   1.52
1966   4.0%   24.0   17.8   2.93%   17.8   1.35
1959   5.4%   17.9   18.8   3.15%   17.9   0.95
1973   4.6%   18.7   18.1   2.67%   18.1   1.03
1961   5.2%   18.4   18.6   3.26%   18.4   0.99
1965   4.2%   23.2   18.7   2.92%   18.7   1.24
1964   4.6%   21.6   18.8   3.00%   18.8   1.15
1962   4.8%   21.1   21.3   2.93%   21.1   0.99
Have fun.

John R.
Post Reply