Regarding Cash Buffer Management

Research on Safe Withdrawal Rates

Moderator: hocus2004

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

Regarding Cash Buffer Management

Post by JWR1945 » Sat Jul 26, 2003 2:20 pm

Regarding Cash Buffer Management

I have been looking into the issue of Cash Buffer Management and the related issue of valuations. Here are some early results. They extend the theory in a familiar direction.

We have established that P/E10 is a much better measure of valuation than we had once thought. Anomalies associated with the earliest period 1881-1920 have masked its value in later periods. There are significant qualitative differences between that period and the years that followed. By focusing on 1921-1980 and retirement portfolios designed to last 30 years, we can gain powerful insights as to why historical results occurred. We can then apply those insights going forward. We can understand much better how our retirement portfolios are doing. This allows us to react as appropriate.

peteyperson identified an especially nasty time period from the past. He credits John Bogle for drawing his attention to it. That period was the fifteen years from 1966-1981. It occurs early in the lives of most of those portfolios that would have failed. Using P/E10 to measure valuations, the 20 years (out of 60) with the highest valuations for 1921-1981 occurred in 1928-1930, 1936-1937, 1956, 1959-1970 and 1972-1973. There are twelve high valuation years in a row (1959-1970). If you add the single year 1971 to the list, you have fifteen dangerous years in a row (1959-1973).

I had hoped to go directly to a cash buffer management approach to increase withdrawal rates (at high levels of safety). That has not yet come about. However, I have been able to extract a number of insights essential to understanding why the historical rates are what they are. That understanding can be applied immediately to monitoring performance. It may provide the genesis towards developing improved strategies.

Using commercial paper caused no harm.

BenSolar had warned me not to expect to increase the calculated safe withdrawal rates by much via changes in allocation percentages. I have now proved him right. My investigation to date has been restricted to tedious, manual calculations on FIRECalc. I have finally put the ability to unzip software onto my computer. I expect to be able to use the Retire Early adaptive calculator in the near future. Using it will be much more efficient. But I have gained some insights already using FIRECalc.

BenSolar had speculated that using commercial paper was a bad choice because of inflation. That turns out to be wrong. In fact, commercial paper did relatively well. I varied withdrawal amounts for a 50% stock allocation and for an 80% stock allocation. I kept the initial balance at $1000. I kept the lifespan at 30 years. I kept expenses at 0.20%. I always used CPI to adjust for inflation. I left the other inputs at their default values.

With a 50% stock allocation and commercial paper, these are the years within 1921-1980 at which portfolios failed:
$38 withdrawal amount, no failures.
$39 withdrawal amount, no failures.
$40 withdrawal amount, 1937 failed.
$41 withdrawal amount, 1937 failed.
$42 withdrawal amount, 1937 failed.
$43 withdrawal amount, 1937 and 1966 failed.
$44 withdrawal amount, 1937 and 1965, 1966 and 1968, 1969 failed.
$45 withdrawal amount, 1936, 1937 and1965, 1966 and 1968, 1969 failed.
$46 withdrawal amount, 1936, 1937 and 1965, 1966, 1967, 1968, 1969 failed.
$47 withdrawal amount, 1929, 1936-1937, 1939-1940, 1964-1969 failed.

With a 50% stock allocation and TIPS at 0.0% interest, these are the years within 1921-1980 at which portfolios failed.
$36 withdrawal amount, no failures.
$37 withdrawal amount, 1966 failed.
$38 withdrawal amount, 1965, 1966 failed.
$39 withdrawal amount, 1965, 1966 and1968, 1969 failed.
$40 withdrawal amount, 1964-1969 failed.
$41 withdrawal amount, 1962, 1964-1969 failed.
$42 withdrawal amount, 1929, 1962, 1964-1969 failed.

Notice that 0.0% TIPS improved results in the Great Depression but did nothing to improve the results in the 1960s. Having commercial paper was better than matching inflation alone.

These are the results with an 80% stock allocation and commercial paper. They are a little bit better.
$40 withdrawal amount, no failures.
$41 withdrawal amount, 1966 failed.
$42 withdrawal amount, 1966 failed.
$43 withdrawal amount, 1965, 1966 and 1968, 1969 failed.
$44 withdrawal amount, 1929 and 1965, 1966 and 1968, 1969 failed.
$45 withdrawal amount, 1929 and 1965, 1966 and 1968, 1969 failed.
$46 withdrawal amount, 1929 and 1965-1969.
$47 withdrawal amount, 1929 and 1937 and 1964-1969 and 1973.
$48 withdrawal amount, 1929 and 1937 and 1964-1969 and 1973.
$49 withdrawal amount, 1929-1930 and 1937 and 1962 and 1964-1969 and 1972-1973.

These are the results with an 80% stock allocation and 0.0% TIPS.
$38 withdrawal amount, no failures.
$39 withdrawal amount, 1966 failed.
$40 withdrawal amount, 1966 failed.
$41 withdrawal amount, 1965, 1966 and 1968, 1969 failed.
$42 withdrawal amount, 1929 and 1965, 1966 and 1968, 1969 failed.
$43 withdrawal amount, 1929 and 1965, 1966 and 1968, 1969 failed.
$44 withdrawal amount, 1929 and 1964-1969 failed.
$45 withdrawal amount, 1929 and 1964-1969 and 1973 failed.
$46 withdrawal amount, 1929 and 1962 and 1964-1969 and 1973 failed.

Once again, the TIPS had an influence during the Great Depression, but they did not help in the 1960s.

There are no increases in withdrawal amounts yet.

The first year with normal valuations after 1959 was 1971. Both 1972 and 1973 had high valuations. In 1974 valuations again returned to their normal range.

Here are the results for the year 1971 with varying lifespans. In every case, a rate that is 0.1% higher had at least one failure. In all cases the stock allocation was 80%. There was 20% commercial paper. The expenses were 0.20%. CPI was the inflation index.

The safe withdrawal rate was 6.1% with 18 years to go. (1959)
The safe withdrawal rate was 5.9% with 19 years to go. (1960)
The safe withdrawal rate was 5.7% with 20 years to go. (1961)
The safe withdrawal rate was 5.6% with 21 years to go. (1962)
The safe withdrawal rate was 5.5% with 22 years to go. (1963)
The safe withdrawal rate was 5.4% with 23 years to go. (1964)
The safe withdrawal rate was 5.3% with 24 years to go. (1965)
The safe withdrawal rate was 5.2% with 25 years to go. (1966)
The safe withdrawal rate was 5.2% with 26 years to go. (1967)
The safe withdrawal rate was 5.1% with 27 years to go. (1968)
The safe withdrawal rate was 5.1% with 28 years to go. (1969)
The safe withdrawal rate was 5.0% with 29 years to go. (1970)
The safe withdrawal rate was 5.0% with 30 years to go. (1971)

The result for 1975 was 7.5% with 15 years to go. All other conditions were the same.

From these results you can determine what your cash buffer must earn for you in order to wait for a good buying opportunity. For example, for a start year 1964, you would draw down your cash buffer account for 7 years until you reached 1971. If you then set your stock allocation to 80% in 1971 and made no other changes, your safe withdrawal rate would be 5.4% of whatever was left as the portfolio balance.

In general, there was no improvement. I made a spot check for buying in 1975 instead of 1971. It was even worse.

Why?

As I searched for ways to improve performance, I kept finding that I needed to have the cash buffer deliver high returns (in real dollars). Even a 3% 10-year TIPS was not good enough. That got me looking at the (real) prices and dividends of the S&P 500 index. I found out something very similar to what I had learned before.

In real dollar terms the dividend amounts were fairly stable. In 1959 the (real) dividend was $10.225 and in 1989 it was $13.678. It got as high a $14.776 in 1967. It fell as low as $11.182 in 1976, but no lower after that. If you started your retirement between 1959 and 1970, you would have something close to 3% as a real dividend yield that grew along with inflation (roughly). You only had to sell enough stocks to add 1% to bring your withdrawal rate up to 4%. Even though you would have to sell some stocks, you would not have to sell many ... even on dips. There were some nasty dips along the way. The (real) index values (truncated) were 323 in 1959, 495 in 1966, 396 in 1971, 235 in 1975, 299 in 1977, 209 in 1982 and 397 in 1989.

Once again, we see the dominance of dividends in determining safe withdrawal rates. You could withdraw an amount of 3% of your initial balance from dividends alone and it would increase to match inflation. The return of capital is minimal at high levels of safety. The age-old concept of living off one's dividends alone and leaving the principal untouched is inherently sound. As a good first approximation, it equals the safe withdrawal rate (except for a small amount received from the return of capital).

An alternative calculation.

This research once again points to dividends as the core element that makes up an alternative strategy for retirement distributions. By focusing on dividends and their safety and prospects for growth, we can add asset classes easily while getting results that we can relate to various studies. There is the issue of the return of capital, but many real life retirees hate the idea of reducing principal. They avoid doing so whenever they can.

This research establishes another basis for estimating safe withdrawal rates at recent (bubble) valuation levels. This approach focuses directly on dividend payouts. An alternative (which I have used) focuses on (ten-year trailing) earnings, which are reasonably well behaved. As we look forward and as dividend amounts come closer to historical percentages of reported earnings (not projected, pro-forma and/or operating earnings), we will be able to relate current conditions more accurately to historical conditions.

Underlying any projection are the dividends and the earnings that support them. There is also the price level if there is to be an enhancement from a return of capital. It is a much smaller factor. The intrinsic value is the critical factor. Valuations are always a critical factor in determining safe withdrawal rates (as a percentage of one's initial balance) because that is what translates the balance to the intrinsic value, whether on a price basis or on an earnings basis.

Have fun.

John R.

Post Reply