A Way of Thinking

Research on Safe Withdrawal Rates

Moderator: hocus2004

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

A Way of Thinking

Post by JWR1945 »

Background

hocus's original question has led us to investigate the safety of retirement portfolios in terms of valuations. That led us first to refine Safe Withdrawal Rate calculations for bubble levels, when valuations extend above the historical range. More recently, we have been able to introduce adjustments for all levels of valuation.

Much of our focus has been on monitoring portfolio performance and identifying failure mechanisms. We have been able to identify some cause and effect relationships. This has enabled us to develop reliable dividend based strategies.

All of this is in the context of some helpful, creative thinking by peteyperson. peteyperson has drawn our attention to intrinsic valuations and cash buffer management. Although we are still at the early stages of our research, we can now design retirement strategies that include additional asset classes and we can monitor portfolio safety successfully. There remain many opportunities for improvement. These include our use of non-cash buffers as well as cash buffers.

What history has taught us

The main reason for retirement portfolio failure is heavy selling at low prices. Looking carefully into failure mechanisms, we have found that the initial dividend yield is only slightly below the Historical Database Rate in times of stress. The Historical Database Rate for any particular year is the highest withdrawal rate that would have survived for the actual sequence of returns that followed. Such times of stress are tightly connected to high valuation levels (as measured by P/E10).

The lowest withdrawal rates that would have been safe in the past can be defined by their initial dividend yields. In a few cases, they have to be adjusted downward because of sharp dividend cuts two or three years later. The actual number increases only slightly (1% or less) if one plans for a final portfolio balance of zero dollars. Dividends have come reasonably close to compensating for inflation, provided that one is talking in terms of many years and provided that one allows for a small number of shares to be sold to fill in any deficiency.

How to think about the problem

It makes sense, conceptually, to divide one's investments into two parts. One component provides a steady, predictable cash income. The other component consists of wildly fluctuating prices and occasional capital gains.

We know that, in times of high valuations, selling stocks can only augment a safe withdrawal amount only slightly (1% or less). We can reduce the negative effects of fluctuating prices with a cash buffer. We can identify an equivalent high valuation price when actual prices are much lower and much closer to intrinsic values. P/E10 provides a suitable scale factor for the S&P 500 index. There may be something better that we have not identified yet. There may be other scale factors that are more suitable for other asset classes.

When prices are in their normal range, one can estimate his safe withdrawal amount from capital gains as an annualized return necessary for his stock to grow from its current price to its equivalent high valuation price over the required lifespan of the portfolio. The exact scale point for the S&P 500 index has not been established. However, if one scales for a P/E10 of 24 (i.e., the equivalent high valuation price = current price*(24 / [P/E10 at the current price] ), one should add 0.74% to 1.36% to that number. The best estimate is to add 1% (with a range of 0.7% to 1.0%). For the upper bound add 1.36%.

To design a retirement strategy, one should identify his income needs first. He should then decide how much of that income should come strictly from steady income sources and how much must come from capital gains. All numbers should include appropriate adjustments for inflation.

The nature of that portion of income that comes from capital gains is that it fluctuates wildly. As a result, you must be able to handle those fluctuations by managing your cash buffer account. The nature of that portion of your income that comes from steady sources is that it does not grow sufficiently to meet all of your needs over an extended period of time. Yet, it may take as long as 15 years before you realize the necessary growth in prices.

The next few steps

The Historical Database Rates were based upon having no flexibility at all. The amount withdrawn remained constant (after adjusting for inflation) and the cash buffer account was never allowed to change. I.e., it was never actually used as a buffer. It is under those conditions that we have come up with the 1% that one should add to income from dividends and interest. Those conditions assume as well that the final balance of one's retirement portfolio should be zero dollars.

It is now appropriate to incorporate peteyperson's thoughts about cash buffer management. We have reached the point that we can scale between a true, intrinsic value and an equivalent high valuation. From the equivalent high valuation we can approximate the results without cash buffer management (i.e., no fluctuations). The next steps are to specify the amount of fluctuation that we can tolerate and then to design a suitable cash buffer.

To introduce other asset classes, it is necessary to think in terms of the potential rewards both in terms of prices (or capital gains) and in terms of interest and dividends. The next step along these lines is to come up with suitable equivalent high valuation formulas for additional asset classes.

Since new asset classes will fluctuate differently (both as to time and to extent), there will be opportunities to take advantage of such differences. Re-balancing does this in part. There may be something else worth doing.

Have fun.

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

Post by JWR1945 »

I have extracted this from one of my posts under the TIPS Mathematics thread on the Newbies board.
http://nofeeboards.com/boards/viewtopic ... 9895#p9895
The real answer is that the REHP study assumes the existence of stocks that did not exist. Repeat, for this is important: the S&P 500, as modeled in the REHP study, did not exist. Nor does it exist today. Nor did the actual S&P 500 stock index come close to anything that was modeled in the REHP.
The REHP study assumes that stocks consistently produce a dividend yield of 3% or more. That was always true in the relevant portion of the historical record. Now that we have been able to identify some basic cause and effect mechanisms, we can focus on that particular fallacy. It caused the REHP study to overestimate the Safe Withdrawal Rate by a factor of two! The REHP number is necessarily wrong. The catchall phrase unless the future is worse than the past actually means unless S&P 500 dividends are less than 3%.

This needs to be understood clearly. The historical sequence approach, which is what the REHP study uses, does not apply today, nor will it apply until dividends return to 3% and higher. This will remain true even when valuations return to in the pre-bubble historical range. Valuations have dipped into the high end of the historical range at times in recent months.

Have fun.

John R.
Post Reply