Investing is a lifelong activity. Yet portfolios and, more broadly, asset allocation models, need to adjust over a client's life span: The portfolio an investor assembles as a 35-year-old is hardly appropriate for a 65-year-old client. It often falls to an advisor to ensure a post-retirement portfolio could handle the market's most violent swings.

One of the most difficult aspects of investing is determining how much risk to take in your portfolio at various stages of life. For instance, portfolio losses during the preretirement accumulation phase are easier to recover from than losses during the post-retirement distribution phase.

The mechanics of an investment portfolio are very different during the distribution phase. Losses are more difficult to recover from because withdrawals only serve to exacerbate the problem.

As shown in the Math of Recovery chart below, a post-retirement distribution portfolio faces a much steeper climb back to break even after a loss than does an accumulation (or buy-and-hold) portfolio. As shown by the shaded boxes, an accumulation portfolio only needs an average annual return of 7.7% to recover from a 20% loss within three years.

But a distribution portfolio, in which money is withdrawn each year, must generate at least a 16.5% average annual return over a three-year period to recover from the same loss. (This distribution portfolio assumes a starting balance of $500,000, an initial withdrawal at the end of the first year of 5% - in this case, $25,000 - and an annual 3% increase in the withdrawal amount.)

The conclusion is quite clear: When building a distribution portfolio for the post-retirement years, it is vitally important to avoid large losses. An investor's post-retirement portfolio must therefore be more loss resistant than the portfolio designed for the early accumulation years. But at the same time, the retirement portfolio must be able to provide sufficient return to preserve and/or grow the portfolio's asset base.

How is this to be done? Examine the performance of various asset allocation models that could be employed to build retirement portfolios.



For this analysis, six asset allocation models in distribution mode were tested to simulate the experience of an investor in retirement, with money being withdrawn from the portfolio (in this case, at the end of each year).

As shown in the 15-Year Retirement Portfolio Survival Test chart on page 67, the asset allocation models were as follows:

* 100% cash (defined as a money-market mutual fund)

* 50% cash/50% bonds

* 60% large U.S. stock, 40% bonds

* 25% each in large U.S. stock, non-U.S. stock, bonds and cash

* A 12-asset diversified model

* 100% large U.S. stock

The time frame of the analysis was the 15-year period from Jan. 1, 1998, to Dec. 31, 2012. The simulated retirement portfolios had a beginning balance of $500,000, although the actual amount is immaterial - the only investor-controlled variables that matter in this analysis are the initial withdrawal rate and the annual cost of living adjustment.

The initial withdrawal rate was set at 5%, or $25,000 at the end of the first year. The annual cost-of-living adjustment was 3% - so the second year's withdrawal was $25,750, the third year's was $26,523, and so on. (This analysis is not intended to endorse or discredit a 5% withdrawal rate.)

Additionally, all the multiasset portfolios were rebalanced back to their percentage allocations at the end of each year.



As shown in the chart, the 100% cash portfolio had a year-end account balance larger than the starting balance of $500,000. After the third year, however, the portfolio's annual ending balance was underwater, or below $500,000 (as illustrated below). By Dec. 31, 2012, the account balance of the all-cash portfolio was $223,941. Over the 15-year period, the internal rate of return - a way to measure the portfolio's returns during a period of withdrawals - was 2.65%.

For a retiree hoping to fund a 25-year (or longer) retirement period, an all-cash asset allocation will likely not get the job done - particularly in a low interest rate environment like the current one.

The next portfolio was a 50% cash/50% bond model, with bond performance represented by the Barclays Capital Aggregate Bond Index. As expected, it outperformed the all-cash portfolio, but was underwater beginning in 2004. The ending account balance in December 2012 was $344,395. The 15-year internal rate of return improved to 4.15%.

Next was the classic 60/40 balanced portfolio: 60% U.S. large-cap stock (as defined by the S&P 500) and 40% U.S. bonds (using the Barclays bond index). This ubiquitous model was underwater at the end of 2002, resurfaced for several years, and then after 2008 again slipped below $500,000. The ending account balance was $433,566, reflecting an internal rate of return of 5.07%.

Also tested was a four-asset model, with 25% each in large-cap U.S. stock, non-U.S. stock (represented by the MSCI EAFE Index), U.S. bonds and cash. It underperformed compared with the 60/40 model and finished with an ending account balance of $383,280 and an internal rate of return of 4.57% - just slightly ahead of the 50% bond/50% cash retirement portfolio.

The next portfolio was a multiasset portfolio consisting of equal allocations to 12 different asset classes (a concept I explain in my book, 7Twelve: A Diversified Investment Portfolio With a Plan). This model includes 8.33% in each of the following categories: large-cap U.S. stock, mid-cap U.S. stock, small-cap value U.S. stock, developed non-U.S. stock, emerging markets non-U.S. stock, REITs, natural resources, commodities, U.S. bonds, TIPS, non-U.S. bonds and cash. This diversified retirement portfolio was slightly underwater on two occasions (1998 and 2002), but then resurfaced and finished the 15-year period with an ending balance of $774,486, with an internal rate of return of 7.73%.

Finally, I simulated an all-stock portfolio, with a 100% allocation to large-cap U.S. stock (or the S&P 500). A 100% stock allocation is rarely recommended as a retirement portfolio, but I included it to provide a point of reference because the S&P 500 is used so commonly as a performance benchmark. The first two years were terrific, but then it began to falter, and by 2002 it was underwater. The account balance resurfaced for two years (2006 and 2007) only to plunge in 2008. In the end, the account balance was $324,447. Its internal rate of return of 3.93% was even below that of a 50% bond/50% cash retirement portfolio.

This analysis clearly demonstrates that diversification is a valuable portfolio construction guideline for distribution portfolios during the post-retirement years. Said more plainly: Diversification makes just as much sense during the post-retirement period as it does during the preretirement accumulation years.

If the last 15 years are any sort of indicator of the future, building broadly diversified retirement portfolios is prudent, logical and beneficial. Simply diversifying among two asset classes (stocks and bonds) is insufficient. The good news is this: With an ever expanding array of mutual funds and ETFs that represent all manner of asset classes, it's never been easier to build diversified portfolios.



Craig L. Israelsen, Ph.D., a Financial Planning contributing writer in Springville, Utah, is an associate professor at Brigham Young University. He is also the author of 7Twelve: A Diversified Investment Portfolio With a Plan.