Bucket Portfolios Faced a Real Stress Test in 2018

By most measures, 2018 was not a great year for investors. Stocks tumbled in the fourth quarter, owing to a grab-bag of worries related to trade, the political impasse in Washington, and slower growth overseas. And while bonds were reliable ballast during that period, they had fallen earlier in 2018, as the Fed's series of interest-rate increases crimped bond prices. In the end, U.S. stocks dropped about 5% for the year, and international stocks fared even worse. The Bloomberg Barclays Aggregate Index managed to remain in the black, but only barely. Higher-risk, higher-yielding bonds generally dropped in value.

But as painful as it was, a year like 2018 also illustrates the point of the Bucket strategy. In years like 2017, when both stocks and bonds notched solid gains, a retiree would be fine with almost any reasonable retirement portfolio and decumulation strategy. Moreover, the cash component that's part of a Bucket approach would be a drag. But lackluster to horrible years like 2018, 2009, and 2002 underscore the virtue of maintaining a cushion of safe assets to use for expenses. Importantly, that cash cushion would prevent a retiree from having to raid depreciated stock or bond holdings to pay the bills. Not only that, but holding a component of cash and short-term bonds might even allow for adding to depressed stock and bond positions during periods of market weakness.

I've been conducting ongoing "stress tests" of the Bucket approach for the past several years, with an eye toward demonstrating the logistics of a Bucket approach for retirees who wish to use a similar strategy for their own portfolios. The year 2018 provided a relatively rare recent example of how the Bucket strategy would work in a year in which long-term positions endured losses. In my simulation, short-term assets, which resulted from rebalancing out of appreciated long-term positions when the market was better, could be used to cover living expenses and to top up depreciated stock positions. (Due to the maintenance regimen outlined below, the safe component of the portfolio had grown larger over the years due to snipping appreciated assets.)

The strategy has also worked quite well over the past 19 years (2000 through 2018), the period covered by my back test. In my simulation, the Aggressive Bucket Portfolio's value stands at nearly $1.9 million, almost $400,000 above its starting value of $1.5 million. And the portfolio has also supplied an additional $1.47 million in cash flows to our hypothetical retirees over that 19-year period. At the same time, it's worth noting that the past decade has been so fortuitous for stocks that almost any sane strategy involving a healthy allocation to equities would have performed reasonably well. You can view the simulation in this spreadsheet (Microsoft Excel required).

A Bucket Overview In a nutshell, the Bucket strategy enables retirees to customize their asset allocations based on their expected spending from the portfolio. Cash-flow needs for the next one to two years (Bucket 1) are parked in the only asset that is guaranteed to not lose money over such a short time frame: cash instruments like money market funds and CDs. Cash-flow needs for the next eight years are earmarked for high-quality bonds (Bucket 2); over the past 25 years, the Bloomberg Barclays Aggregate Index has generated positive returns in 100% of rolling three-year periods. The remainder of the portfolio (Bucket 3) goes into stocks and higher-risk assets like junk bonds; because the retiree holds at least 10 years' worth of cash flow needs in Buckets 1 and 2, the higher-risk assets aren't so risky after all. The S&P 500 has generated positive returns in more than 90% of rolling 10-year periods over the past 25 years.

In the stress test of my aggressive mutual fund portfolio, I assumed a $1.5 million portfolio with a 4% initial withdrawal, translating into $60,000 in year 1 of retirement (2000 in my simulation). That withdrawal amount is then adjusted annually to account for inflation, but the retiree foregoes the inflation adjustment in years in which the portfolio loses value (like 2018).

The starting portfolio used in the simulation is as follows. Note that there are variations between these holdings and the ones featured in the actual Aggressive Mutual Fund Bucket portfolio. In some cases I had to use proxies in the simulation to account for varying inception dates and strategy changes. For example, Vanguard Dividend Appreciation VDADX is the core equity holding in my model portfolio, but it wasn't around in 2000, the inception date for the stress test so I used T. Rowe Price Equity Income PRFDX instead. I've also made some changes to the holdings in my actual model portfolios since their inception date. For example, I dropped commodities exposure in 2017, and in 2018 I swapped in American Funds International Growth and Income IGIFX in place of Harbor International HAINX. I also swapped in Fidelity Short-Term Bond FSHBX in place of T. Rowe Price Short-Term Bond PRWBX when the latter was downgraded to a Neutral rating.

Bucket 1: $120,000 $120,000: Cash

Bucket 2: $480,000

$130,000: T. Rowe Price Short-Term Bond

$150,000: Harbor Bond HABDX $100,000: PIMCO Real Return PRRIX $100,000: Vanguard Wellesley Income VWINX

Bucket 3: $900,000

$400,000: T. Rowe Price Equity Income

$200,000: Harbor International

$100,000: Vanguard Total Stock Market IdexVanguard Total Stock Market Index VTSMX $125,000: Loomis Sayles Bond LSBRX $75,000: PIMCO Commodity Real Return Strategy PCRAX

Performance Update Just as 2018 was a poor year for investors, so it was for the equity-heavy portfolio. All but the cash and short-term bond fund lost value for the year, and Harbor International posted double-digit losses. I used T. Rowe Price Equity Income as a proxy for my actual portfolio holding, Vanguard Dividend Appreciation, but the T. Rowe fund underperformed last year and in 2016 and 2017, too.

Because the portfolio didn't gain in value last year, I assumed the same withdrawal amount as in 2017. I withdrew money from the short-term bond fund, which had grown large due to the maintenance regimen outlined below, to meet the hypothetical retiree's cash-flow needs and to top up depreciated positions back to their starting values.

I'll concede that the maintenance regimen I've been employing throughout the stress tests is idiosyncratic. For one thing, it calls for rebalancing positions that exceed 110% of their starting values. That's hardly a conventional rebalancing regimen, but I found that setting rebalancing thresholds at the security, rather than asset class, level facilitated regular cash-flow extraction.

Moreover, the rebalancing system I used called for adding excess portfolio amounts to the short-term bond fund, once cash-flow needs have been met and all positions have been topped up to their original sizes. The net effect of that strategy over the past decade, as stocks have appreciated, is that the portfolio has becomes progressively more conservative. Of course, our hypothetical retired couple would also be 19 years older; assuming they were 65 in 2000, they'd be 84 today. Given that and a not-cheap equity market, a more conservative asset allocation seems reasonable. At the same time, a lighter equity allocation--and the constrained long-term growth potential that it implies may or may not be what every retiree seeks.

Maintenance Regimen Here are the details on the portfolio maintenance regimen I employed in the "stress test"; you can also see it reflected in the portfolio simulation.

• Withdraw 4% from the portfolio in year 1 (2000) of retirement. Inflation-adjust that dollar amount annually, but forgo inflation adjustments in years in which the portfolio loses value.
• Reinvest all dividends and capital gains.
• Refill Bucket 1 using rebalancing proceeds. Portfolio rebalanced annually if position sizes exceed 110% of their original size. (This leads to more rebalancing opportunities than would be the case with classic rebalancing.)
• If rebalancing proceeds are less than annual living expenses, pull the remainder of needed cash flow from T. Rowe Price Short-Term Bond in Bucket 2.
• If rebalancing proceeds are greater than annual living expenses, move any remaining proceeds into positions that have declined in value since inception.
• If rebalancing proceeds are greater than cash needs and long-term positions are at original size, add additional monies to cash.
• If cash holdings exceed three years' worth of living expenses and long-term positions are at original size, move additional monies to short-term bond fund.