Joined: 26 Nov 2002
Location: Crestview, Florida
|Posted: Sun Feb 06, 2005 11:52 am Post subject: Accumulation Overview
I have made several investigations into the accumulation phase of investing. I have looked at combinations of two portfolios. The first uses dollar cost averaging into a stock-only portfolio. The other switches allocations of stocks and TIPS with a 2% interest rate as a function of P/E10, which is a measure of valuation. [There are additional conditions with commercial paper or ibonds for the non-stock component.]
My original investigation showed the following:
Accumulation is Different dated Tue Oct 19, 2004.
1) When starting out with an initial balance of zero, it is better to dollar cost average into an all-stock portfolio for 30 years than it is to dollar cost average into a portfolio that switches allocations for 30 years.
Switching improved the worst case results only a small amount while taking away the upside. Frequently, the upside potential was large.
Final balances were almost entirely independent of the percentage earnings yield 100E10/P.
2) When starting out with an initial balance of zero, it is usually better to dollar cost average into an all-stock portfolio for 15 years than it is to dollar cost average into a portfolio that switches allocations for 15 years.
Switching allocations produces a non-trivial improvement when overall stock returns are low, but with a serious penalty on the upside. The entire decade of the 1960s produced better results with switching than without. The decade from 1925-1934 favored switching as well.
In general, switching allocations produces better results when the percentage earnings yields [at the very start of a historical sequence] is below 6% (i.e., P/E10s of 17 and higher).
The 15-Year results favor switching at today's valuations, but not at the valuations typically found in the historical record.
Later, I broke the 30-year period into two 15-year periods. I used dollar cost averaging to add funds only during the initial 15-year period. I left the balance untouched (that is, I made neither contributions nor withdrawals) during the remaining 15 years.
Accumulation is Different dated Tue Oct 19, 2004.
1) I looked at investing entirely in stocks for both periods. This approach produced the worst case result for the modern era (1923-1972).
2) I looked at investing into a portfolio that switched allocations during the first 15 years and then allocated everything into an all-stock portfolio for the final 15 years.
Switching provided a small improvement in the modern era about one-half of the time. It caused a large reduction one-fourth of the time.
3) I reversed the sequence of events. I made deposits into an all-stock account for the initial 15 years followed by switching allocations during the final 15 years.
The upside was limited once again. But this time worse case balances increased enough to justify sacrificing some of the upside potential.
Later, I examined initial balances other than zero.
Some Accumulation Multipliers and Examples dated Tue Nov 09, 2004.
Additional tables for the accumulation stage dated Tue Dec 28, 2004.
1) Under these conditions, what is best for dollar cost averaging differs from what is best for a balance that is left untouched.
2) I produced do it yourself type tables and two examples. I included instructions for following an initial 15-year period that includes new deposits by another 15 years without any further deposits.
3) I switched allocations with the entire initial balance. I dollar cost averaged starting from a balance of zero into an all-stock portfolio.
4) I did not examine making fine grain adjustments. Such an adjustment might be to abandon dollar cost averaging into the all-stock portfolio when the initial value of P/E10 exceeds a high threshold.
5) The best and worst results for a combined portfolio can differ from what is best and worst for the individual components. [The individual components are the initial balance and the deposits made for dollar cost averaging.]
I verified that the algorithm used for switching allocations is a good choice.
Switching during Accumulation dated Mon Dec 27, 2004.
I stopped my investigation early. The best switching algorithm for making withdrawals during distribution is not far from being the best for portfolios that are left untouched.
I did not continue my investigation long enough to prove that they are identical.
Most recently, I have provided numbers to assist in transition planning.
Numbers for Transition Planning dated Sat Feb 05, 2005.
These numbers show the reward and risk as a function of valuations (in the form of 100E10/P, the percentage earnings yield of the S&P500) for a variety of portfolio allocations at 5 years and at 10 years before retirement.
In terms of what is most likely to happen 5 years from now and using today's valuations, investing entirely in TIPS (at a 2% interest rate) is the best choice. It produces a 10% gain with certainty. With a 20% stock allocation, you are highly likely to retain most of your initial balance even under worst case conditions. Your most likely outcome is a gain of 7%. With a 50% stock allocation, you can lose up to 40% of your initial balance under worst case conditions. Similarly, you can gain up to 40% under best case conditions. Your most likely outcome is just above breaking even.
In terms of what is most likely to happen 10 years from now and using today's valuations, investing entirely in TIPS (at a 2% interest rate) is an excellent choice. It produces a 20%+ gain with certainty. With a 20% stock allocation, you are highly likely to add to your initial balance even under worst case conditions. Your most likely outcome is a gain of 20%+, and just a little better than from TIPS by themselves. With a 50% stock allocation, you can lose almost 30% of your initial balance under worst case conditions. You can gain 90% under best case conditions. Your most likely outcome is a gain of 20%+, but slightly less than from TIPS by themselves.
It takes values of P/E10 around 16 or 17 or lower before it is safe enough to make a clear-cut, compelling argument in favor of holding 50% stocks just before retirement. It takes values of P/E10 below 12 before arguments in favor of higher stock allocations just before retirement become clear-cut and compelling.