Safe Withdrawal Rate with Momentum/Trend-Following

November 12, 2025 – Hello, readers! Welcome to another installment of the Safe Withdrawal Rate Series. Please see this landing page for an introduction to the Series and a summary of all the other parts so far. After a long hiatus from writing due to my busy travel schedule during the summer and lots of other commitments, I’ve found my groove again and put together something that has been on my mind for many years: Is there an asset allocation strategy that could have improved historical safe withdrawal rates? Specifically, could we devise an asset allocation strategy that shifts weights between different asset classes in a way to improve investment results? Of course, that’s easier said than done, but there are some interesting ideas out there. One such approach is to tactically shift asset class weights based on asset return momentum. Some people also refer to this flavor as “Trend-Following.” If you want to sound really techy and fancy, you’d also call this “Tactical Asset Allocation” (TAA), “Managed Futures,” or “Commodity Trading Advisers” (CTA) strategies; however, these three terms often encompass many other dynamic asset allocation strategies, not just momentum.

In any case, maybe a momentum strategy can help us avoid some of the worst historical asset market disasters if we could sell equities early enough during a bear market. How much Sequence Risk could we eliminate? By how much can we raise our safe withdrawal rate if we could have reliably avoided some of the worst historical asset market disasters? Let’s take a look..

Before trying Momentum, what else could hedge Sequence Risk?

Before we even jump into momentum, is there anything else worth pursuing? If you’ve followed my Safe Withdrawal Rate Series, you’ll remember that I’ve taken a very pessimistic view on the efficacy of all the proposed asset allocation tricks to improve safe withdrawal results:

What most of these strategies have in common is that they primarily pursue a fixed asset allocation with slightly different weights than the traditional 60/40 or 75/25, plus a few additional asset classes. Well, there is no free lunch: Only total returns matter, and a higher dividend yield lowers your expected price return. Stock-picking styles like Small-Cap, Value, and Small-Cap Value are no longer a reliable source of alpha, now that everyone is aware of them. Risk Parity is such complete charlatanry that I’m not even going into the weeds here. Stay tuned for a future post.

So, consistently holding the existing asset classes available to the average retail investor is not that fruitful in expanding the investment horizon. Then, one new route we should explore is the “smart” way of shifting between the different asset classes. One such approach is momentum.

Why would asset market momentum work?

In a nutshell, asset market momentum looks more promising than the previously proposed solutions from the mutual fund salespeople because…

1. Asset returns depend on economic regimes, such as recessions versus expansions, inflationary versus deflationary regimes, tight versus loose monetary policy, as well as commodity cycles of abundance versus scarcity, which can take time to return to normal.
2. Economic regimes persist for extended periods. For example, the probability of being in a recession next month depends on the current economic state: you’re more likely to be in a recession next month if we’re in a downturn this month. The Federal Reserve often goes through a whole sequence of rate hikes or cuts. There are production and harvest cycles in many commodities. And so on.

Let me justify both points really quickly with some empirical evidence. First, average asset returns vary significantly across different economic regimes, specifically between NBER expansions and recessions, as shown in the table below.

Nominal Asset Class Returns in % During Economic Regimes. 1926-2025.

• Equities tend to perform well during expansions, offering high returns and low risk. However, they suffer during recessions, experiencing negative returns and almost twice the volatility compared to an expansion.
• Intermediate U.S. Treasury bonds do very well during recessions. But they still have decent returns even during expansions. In fact, even during expansions, they have a tiny edge above cash/T-bill returns (3.87% vs. 3.29%).
• A comprehensive index, such as the GSCI index, which encompasses all major commodities, follows a similar pattern to equities, yielding significant returns during expansions. But poor returns and high risk during recessions.
• The Gold spot price performs well across all economic regimes, but it has a particularly attractive return profile during recessions. Not as good as Treasury Bonds, though. Also, notice that Gold is a member of the GSCI Commodity Index, so it’s captured in the Commodities return series. But gold certainly has very different return characteristics from the others!
• Cash, such as 3M T-bills or related assets like money market accounts, performs well under all regimes. Even slightly better during recessions, which would make it an ideal risk-off/safe-haven asset during economic turmoil. Gold indeed performs somewhat better on average during a recession (7.27% vs. 4.08%), but at the cost of significantly higher volatility.

I also like to include a few more fun facts. Most asset class returns correlate not just with the current month’s economic regime. The correlations are even stronger if we offset the economic regime by a sufficient number of months. For example, equity returns correlate best with the economic status four months in the future, as shown in the table below. That’s perfectly intuitive: the equity market leads the economic cycle. Commodities by about 1-2 months, gold by one month. Also noteworthy is that the short-term return, closely tied to the Federal Reserve policy rate, leads the cycle by the longest, approximately 6 months.

Asset class return correlation with the NBER expansion indicator (1.0 for expansion, 0.0 for contraction) at various leading months. 1926-2025.

Next, let’s look at some evidence of U.S. economic momentum. If we examine the NBER expansions versus contractions over time, we find that over the last 100 years, we have been in a recession about 16.8% of the time; however, the recessions have been clustered. Conditional on being in an expansion in one month, we had a 98.4% chance of continuing the expansion next month. Likewise, if we were in a recession in one month, we’d stay in a recession with 92%. The “escape probabilities,” i.e., the chance of changing the economic regime, are 8% from a recession, implying a roughly 13-month average recession length, and only 1.6% from an expansion, implying a roughly 62-month average expansion length. And just FYI, since 1970, the recessions have been slightly shorter and the expansions have been significantly longer, more than 80 months.

Recession vs. Expansion probabilities in %, conditional on last month’s regime as well as unconditional. 1925-2026.

But isn’t the market efficient?

One objection I can already predict: if there is this predictable and reliable market return pattern, why isn’t this all priced in immediately? My response: Nobody knows in real-time whether we’re in a recession or an expansion. It’s also true that around the turning points, the market moves very rapidly, and prices in a significant portion of the expected outperformance of the new expansion or the underperformance of the impending recession. But certainly not all. For example, after the stock market trough in early 2009, the S&P 500 rallied by over 50% (from February 2009 to February 2010). However, it still left enough runway for the massive rally, which lasted until February 2020. Therefore, the positive news of a new and developing expansion and bull market will filter in gradually over time, which is one of the main reasons for asset return momentum.

Designing a Momentum Strategy

There isn’t one single way to design a momentum strategy. I will propose one that makes sense to me. It will not be the most effective momentum strategy available. However, keep in mind that alternative strategies with numerous bells and whistles are likely to suffer from in-sample bias; that is, someone may go to great lengths to maximize performance, making the backtest appear good, but at the risk of overfitting and potentially creating worse results out of sample.

Momentum assets

For the time being, I propose a momentum asset allocation strategy that utilizes four major asset classes: Equities (S&P 500 Total Return Index), U.S. Intermediate-Term Treasury Bonds (10-year Treasury Benchmark Total Return Index), Gold Spot Price, and 3-month T-bills. I will simulate returns since January 1871 through June 2025. Also, note that I use the returns from my SWR Google Sheet. See Part 28 of the series for the link to the sheet and for a guide. Notice that I use only monthly return data. In recent decades, we certainly have all series in daily frequency, and we’d be able to design a much more intricate momentum signal. But early in the sample, I only have monthly data, so I stick with the monthly frequency.

Momentum assumptions

There is no shorting allowed. Clearly, we could improve the results by allowing short positions of those asset classes that find themselves in a negative momentum spiral. Professional investors, like CTAs/Commodity Trading Advisors, will create the desired long and short positions through futures. However, the average retail investor will be hard-pressed to effectively and affordably short the major asset classes. Yes, there are some short-exposure ETFs, but their expense ratios and other costs appear excessive; more on that later.

I target a base asset allocation of 70% Equities, 20% Bonds, and 10% Gold in case all three asset classes display positive momentum. Any remaining portfolio weight not allocated to the three major asset classes is in the Cash bucket.

I calculate the momentum at three different horizons: 8, 9, and 10 months. Why those horizons? That’s precisely what experienced momentum traders tend to target. Any shorter horizon will create too many false alarms, and any longer averages will cause the momentum signal to miss the boat around the turning points. I also use multiple horizons to increase robustness and potentially spread out trades around the turning points.

I also calculated the momentum signal in two different ways. Notice that we can compare the most recent index level to the rolling average of index levels. Or, we can compare today’s index vs. the index N months ago. There’s also one additional twist: for the equity momentum signal, I compare the average of the latest two monthly index levels with the rolling n-month average. This is a bit of inside baseball, but equity returns are very slightly negatively correlated between two adjacent months (i.e., negative serial correlation). To smooth out that short-term volatility, I take the average index level over the last two months. Think of that as the monthly frequency equivalent of a 50-day over 200-day rolling average crossover signal.

Momentum Formulas.

Next, there is an additional design choice: Do I take the raw indices or the excess return over the cash return series? Tough choice! So, I calculate both. On the one hand, many momentum traders simply look at the raw index series. But on the other hand, I can also justify that if Cash is a fallback asset class, then we should use the excess returns over cash as our target series.

Summary so far: for each asset class, I generate 12=3x2x2 momentum signals, which is the product of three momentum horizons (8,9,10 months), two momentum formals (rolling average crossover versus the T-N to T asset return), and two index versions (total return versus total return in excess of the cash return series).

How to translate the momentum signal into asset class weights

I calculate each of the 36 momentum signals, 12 for each asset class, and assign 1.0 for a positive momentum raw signal and 0 for a negative raw signal. Then average over each asset class to get a raw momentum signal between 0 and 1 for equities, bonds, and gold, respectively.

Another crucial assumption is how the negative momentum weights are redistributed to the other asset classes. The most straightforward approach would be to transfer all unused equity, bond, and gold weight directly into the cash bucket. But that would be pretty inefficient. Why shift into the cash bucket if there might be other, more productive and profitable assets with positive momentum? So, I propose the following sequential reshuffling of the unused asset class weights: The weight not used by gold is added to the equity base weight. The part not used by equities is added to the bond base weight. Finally, the portion not used by bonds goes to the cash bucket. Here’s an example: Imagine that gold has a momentum signal of 0.25, equities have 0.75, and bonds have 0.50. Here would be the asset allocation weight calculation:

• Gold gets a weight of 2.5%, which is 0.25 times the base weight of 10%.
• Equities get a new, adjusted base weight of 70% plus the leftover from the gold allocation (7.5%), so 77.5%. Since the momentum signal is 0.75, equities get 58.125%. The remaining 19.375% carries over to bonds.
• Bonds receive a new base weight of 20% plus the 19.375% from equities, for a total of 39.375%. However, the momentum signal is only 0.50, so only 19.6875% of the allocation goes to bonds.
• The remaining 19.6875% shifts to cash.

Other details

I assume the following annual expense ratios for the different asset classes:

• Equities 0.03% (e.g., iShares IVV or Vanguard VOO)
• Bonds 0.15% (e.g., iShares IEF)
• Gold 0.09% (e.g., iShares IAUM)
• Cash 0.09% (e.g., Shares SGOV)

Furthermore, I assume that each transaction incurs a 0.03% drag on performance for each asset class due to commissions, bid-ask spreads, and other related costs. For example, shifting 40% of the portfolio from equities to bonds will trigger a fee of 2 × 0.4 × 0.03% = 0.024%, i.e., the t-cost hits both the bond and equity sides, hence the 2x. I was concerned that turnover costs would undermine the returns in the simulations, but I was pleasantly surprised: the drag from turnover was only about 0.09% per annum. Well, that’s assuming today’s financial technology and low or even zero-cost commissions at Fidelity and Vanguard. An actual retiree in 1929 would have paid more in transaction costs back then! Therefore, as always, we should acknowledge that these simulations involve a thought experiment that utilizes today’s financial technology, while assuming that some historical return sequences may repeat in the future. Under no circumstances would I claim that a 1929 retiree would have gotten exposure to the US stock market with just a 0.03% annual expense ratio and 0.03% transaction costs back then!

Momentum Signal Time Series

Let’s plot the asset class weights during a few interesting time periods. Let’s start with the period from 1925 to 1945. The strategy successfully shifted out of equities in late 1929, with only a few false alarms in between. For the most part, we were heavily invested in Treasury bonds, which performed very well during this deflationary period. Then, they returned to equities in 1933 to capitalize on the subsequent recovery. The reason Momentum did so well was that the drawdown was long and deep enough that even with a bit of a delay, you still would have stayed out of the equity market for much of the economic and financial malaise. Also, notice that gold wouldn’t have been a very prominent part of the portfolio during this period.

Asset Class Weights: 1925-1945.

Next, let’s examine the period from 1965 to 1985. There were several periods when you shifted out of equities and straight into cash because bonds were also suffering during this inflationary period. Gold played a more prominent role than in the 1930s. Momentum was a useful tool during this period!

Asset Class Weights: 1965-1985.

Finally, let’s examine the period from 2000 to 2025. We started with the backend of the 1990s equity bull market but quickly shifted out of stocks and into Treasury bonds during the dot-com bust. Treasuries performed well because you captured the long path down in yields, which generated excellent bond returns from the duration effect. In 2003, you returned to equities and rode that positive momentum until the onset of the Global Financial Crisis (GFC). Momentum also navigated the Global Financial Crisis very well, returning to Treasury bonds until 2009. The impressive equity bull market that followed 2009 was clearly interrupted here and there with some false alarms; i.e., we briefly dropped equities only to jump back in again after a few months. We also did quite well during the pandemic (only dropped equities very briefly) and the 2022 bear market, where we nimbly shifted into all cash because all risky assets hurt back then. Sweet!

Gold had a relatively consistent allocation of 10%, with only a few exceptions, which proved to be the right decision. This was because gold was one of the best-performing assets since 2000, outperforming the S&P 500.

Asset Class Weights: 2000-2025.

Momentum Strategy Simulation Results

The momentum strategy results look pretty promising. Here’s the cumulative performance, adjusted for US CPI inflation. I plot three return series:

1. A portfolio with static weights of 70% equities, 20% bonds, 10% gold, and no cash allocation, which are the base weights targeted in the Momentum Strategy if all three signals are positive.
2. A portfolio with static weights equal to the overall average weights of the momentum strategy, which was roughly 52% equities, 27% bonds, 3% gold, and 18% Cash.
3. The momentum strategy with the dynamic tactical asset weights.

I also plot the excess performance of the Momentum strategy over the two static strategies with fixed weights. Notice that the vertical scale is in logs to make the growth rates over time easier to see. The Momentum results look really impressive: It left the other two strategies in the dust! Granted, most of the outperformance relative to the 70/20/10/0 portfolio occurred during the period from 1871 to about 1940. However, the path of the momentum strategy was also much smoother post-World War II. Also, the higher average weight in equities helped the 70/20/10/0 portfolio. The asset allocation alpha has remained relatively consistent in the post-World War II period, as indicated by the purple line.

Real Cumulative Returns: 1871-2025.

In the following chart, I plot the drawdowns of CPI-adjusted returns of the three asset allocations. The 70/20/10/0 strategy typically has the deepest drawdowns, while the momentum strategy usually has very shallow drawdowns. Well, not all the time! One fly in the Momentum strategy ointment: it had worse drawdowns during the 1940s. It’s understandable because asset returns during wartime may swing too rapidly for a momentum strategy to capture correctly. So, the WW2 performance is a bit of a downer (though, surprisingly, Momentum did well during WW1, which is a bit of a surprise). This will also have some unpleasant consequences for the SWR analysis later. The momentum strategy worked remarkably well over the last 30+ years, only barely exceeding 10% drawdowns, while the static allocation strategies suffered significant drawdowns of 30% or more during the dot-com and Global Financial Crisis. Notably, momentum performed well during the most recent downturn in 2022, with only a 12% drawdown, while the static allocations experienced drawdowns of 20% or more. This is impressive, considering that the 2022 mini bear market was very short-lived, which is not exactly where an 8- to 10-month momentum strategy is guaranteed to work well.

Real Drawdowns: 1871-2025.

In the table below, I present the detailed return stats. There are two panels: the top displays the nominal returns, and the bottom shows the real, CPI-adjusted returns. Momentum outperformed all individual asset classes and the strategic asset allocations, including a 75/25 portfolio, which is my preferred retirement asset allocation at present. Momentum generates average returns that are about in line with equities, but with vastly smaller volatility and drawdowns.

Simulated return stats: 1871 to 2025.

Here is the same table for the last 30 years (January 1995 – September 2025). Notably, since 1995, the Momentum strategy has achieved an impressive 0.95 Sharpe Ratio using nominal returns. Qualitatively, we obtain the same results: Momentum performs well, within 100 basis points of annual equity returns, but with almost 50% less risk. Also, the drawdowns were very manageable. Note that when computing the “strategic” or fixed asset allocation, I use the average weights from this 30-year period, which came out to be 57% equities, 25% bonds, 6% gold, and 11% cash (totals don’t add up to 100% due to rounding).

Simulated return stats: 1995 to 2025.

Testing the Momentum Strategy in the SWR Google Sheet

How do I add the momentum signal returns to my Google Sheet? That’s amazingly easy: In my sheet, I have space for two “Custom Series” in the Asset Return tab. I currently store the momentum series in the Custom Series 1 column, which grabs the cumulative returns of my simulated series in column K in the “Asset Returns” tab.

I simulate four different asset allocations:

1. My usual baseline is 75% equities and 25% Intermediate Bonds.
2. The base weights form the momentum strategy, i.e., 70% equities, 20% bonds, 10% gold, and thus 0% cash.
3. The Momentum strategy. I assume that 100% of the portfolio follows this approach.
4. Half of the portfolio follows a momentum strategy, while the remaining 50% remains invested in equities. I use this as a proxy for investors who may have a constraint on how much of their assets they can tactically shift around. Think of an investor who prefers to maintain a floor of 50% of their portfolio in equities in a taxable account but can allocate the remaining assets across tax-advantaged accounts without incurring significant capital gains taxes.

Furthermore, I study two different retirement horizons: one with the standard 360 months horizon (e.g., a traditional retiree) and one with 600 months, i.e., the typical early retiree. Let’s assume our retirees are OK with asset depletion, i.e., a final portfolio value of $0.

In the main tab, I would enter the 100% Momentum strategy as shown in the screenshot below. Note that I set the expense ratio to 0.00% because all the momentum returns are already net of expense ratios and trading costs.

How to simulate the Momentum Strategy in the SWR Google Sheet.

What effect does momentum have on the fail-safe withdrawal rates? As usual, I report the failsafe withdrawal rates by decade, so we can get a sense of how the different asset allocation strategies performed during the historical worst-case scenarios.

Let’s start with the 360-month retirement; see the chart and table below. Just as a refresher, even the baseline 75/25 portfolio (or any other Stock/Bond or Stock/Bond/Cash combination) had a safe withdrawal rate of slightly less than 4%; here it’s 3.82%. Not 4% as often falsely claimed. So, the 4% Rule is not 100% but only about 99% safe. See Part 1 of my series. However, the good news is that 4% Rule failures are rare. Only in the most outrageous financial crashes (1901-1904, 1929-1933, 1968-1982) would the 4% Rule have failed.

The Momentum Strategy performed remarkably well during the 1900s and 1920s, achieving a SWR of over 5%. Even in the 1960s, you reached 4.79%. As mentioned above, there is one fly in the ointment, namely the 1930s and 1940s. If you had retired right before the 1937 market peak, your SWR would have been only 4.07%. That’s the overall worst-case retirement cohort for the Momentum Strategy. The reason is that between 1938 and 1945, the Momentum strategy completely messed up the equity timing, as some of the large drops lasted only one or two months. That’s too short to capture with Momentum. Oh, well, you can’t win all the time!

Failsafe Withdrawal Rates by Decade. 30y Horizon.

Notably, the 50% Stocks + 50% Momentum mix performed well. You didn’t quite reach the 5%+ in the 1900s and 1920s, but you also avoided some of the equity market whipsaw in the 1930s and 1940s. If we examine the overall worst-case scenario for the 50/50 mix strategy, it remains the August 1929 cohort, but with a significantly improved SWR of 4.24% compared to 3.84% for the naive 75/25 portfolio. That’s more than a 10% increase in the retirement budget. Nothing to sneeze at!

Next, let’s move on to the 50-year horizon. Notice that I include the decades only up to the 1970s to capture the 1972 scenario. But the 1980s and 1990s don’t have a full 50 years of return data. As always, extending the horizon lowers my withdrawal rates. The overall worst-case scenario for the 75/25 portfolio was now in the 1900s (3.35%), closely followed by the 1920s (3.39%), 1960s (3.43%), and the 1910s (3.56%). In each of these decades, a Momentum strategy would have vastly improved your results and afforded you a SWR significantly above 4%. Unfortunately, we have that unpleasant 1937 cohort again, which was whipsawed during the 1938-1945 equity volatility. Aside from that one little mishap, the Momentum strategy trounced the 75/25 and the 70/20/10. Just like in the 30-year case, mixing the Momentum strategy half-half with an all-equity portfolio would have done very well in the 1930s and 1940s, though you would have slightly lagged behind the exceptionally strong performance in the other decades. So, the inflexibility due to the taxable account could have been a blessing in disguise: The half-half strategy had the overall best failsafe withdrawal rate: 3.86% relative to the 3.35% in the 75/25 portfolio. That’s a 15% improvement in the annual retirement budget.

Failsafe Withdrawal Rates by Decade. 50y Horizon.

What about Momentum ETFs?

Implementing and regularly monitoring this momentum strategy, while not exactly rocket science, may still be too much for some investors. So, are there any ETFs that can perform the same task? Note that an ETF would also have a significant tax advantage: you’d keep the realized capital gains within the ETF wrapper and only owe capital gains taxes when you sell the ETF itself. Though the ETF would still regularly pay out taxable dividends, of course. Moreover, ETFs can likely do more advanced momentum strategies, for example, shorting some indexes or commodities through futures. But will a Momentum ETF do better? I have my doubts! Several ETF offerings pursue so-called managed futures strategies, i.e., run a momentum strategy implemented through futures contracts:

The last one has only a very short history, but the other three at least go back to the pre-pandemic era. Let me calculate the return stats since at least the DBMF ETF, which started in 2019. Additionally, to address the brief history of the KMLM, I have filled in its early return data for 2019 and 2020 using the DBMF returns. Here are the return stats, all in CPI-adjusted real returns. As standalone strategies, most, if not all, of the Managed Futures funds were subpar. The WTMF, FMF, and KMLM returns were just barely above inflation. Even the DBMF produced only 3.35% annualized real returns, which is a bit thin in retirement. I should also point out that I’m not cherry-picking the time period because the WTMF and FMF have returns going back further; however, the results were even worse before 2019.

Return Stats of selected Managed Futures ETFs: 5/2019-9/2025. All returns are real, CPI-adjusted. *For KMLM returns in 2019-2020, I use the DBMF returns

Here’s a chart with cumulative real returns, i.e., adjusted for U.S. CPI inflation. One quick comment: I hope you recognize why I took the liberty of backfilling the KMLM returns with the DBMF returns in 2019-2020; KMLM and DBMF seem to have very similar trend-following models. They are highly correlated during the period when both have data, albeit the KMLM appears to have a consistent drag relative to the DBMF, so the KMLM returns might have looked even worse if they had started earlier.

Real Total Return of selected Managed Futures ETFs: 5/2019-9/2025. *For KMLM returns in 2019-2020, I use the DBMF returns

In any case, DBMF and KMLM indeed performed admirably well until the Fall of 2022; they must have quickly identified the 2022 bear market and shorted most asset classes with negative momentum, hence the strong positive performance in the spring of 2022. Of course, the price you pay when shorting risky assets is that around the turning point, you may face some serious pain. All the impressive gains from shorting assets during the 2022 bear market evaporated again within a few months when the market turned around more quickly than the trend-following models could reverse their positions from negative to positive. The Lord Giveth, The Lord Taketh! Neither DBMF nor KMLM is even close to its prior high. WTMF and FMF also look pathetic. In contrast, the ERN Momentum strategy just chugs along from one high to the next. All this, despite (or perhaps because of) being a long-only strategy, with no shorting allowed.

And, finally, the drawdowns. KMLM is a complete clown show. DBMF also had close to 20% drawdowns. That’s less than the equity market in 2022, but not much. FMF had only modest drawdowns, but it’s also the ETF with the lowest average returns.

Real Drawdowns of selected Managed Futures ETFs: 5/2019-9/2025. *For KMLM returns in 2019-2020, I use the DBMF returns

So, my verdict here: Stay away from the Managed Futures ETFs. They only enrich the financial services industry. People who push these funds are charlatans. If you like this momentum approach, you’d likely have to implement it yourself. If you absolutely want a momentum strategy, the best option currently available may be the DBMF, but it has not yielded particularly impressive returns in absolute terms. It only outperformed three absolutely terrible competitors.

Conclusion

There you have it. A momentum strategy, while not a perfect hedge for Sequence Risk, could be the key to a slightly safer retirement. Even a relatively naive and mundane equity/bond/gold momentum strategy outperformed a static asset allocation, not just in plain buy-and-hold return stats over the last 150+ years, but especially in safe withdrawal rate simulations for most of the historical retirement cohorts. The reason is that you achieve average returns almost as high as an equity portfolio, but with significantly lower volatility and far lower drawdowns. That’s the secret sauce you need to mitigate Sequence of Returns Risk.

What I found most noteworthy is that the momentum signal has consistently helped you over the last 150 years. Sometimes, you test a strategy and it works well during the first half, but then fizzles during the second half. Not so for momentum. It has been remarkably consistent across all major asset market cycles, particularly in the equity market volatility of the last thirty years: the Dot-Com bubble, the Global Financial Crisis, the Pandemic, and even the relatively short and shallow, garden-variety bear market in 2022. If anything, Momentum became more reliable recently. For full disclosure, the major risk would be a repeat of the jagged equity performance of the 1938-1945 period. However, if we have confidence that future recessions and bear markets are more in line with those of the last thirty years, you might find this momentum approach appealing.

Would I switch my portfolio to this momentum strategy now? I will certainly monitor the momentum signals in the future. I wouldn’t liquidate any of my Fidelity equity mutual funds in taxable accounts, but I may shift around some funds in tax-advantaged retirement accounts. If there is demand, I may even update the momentum signals regularly and post them either on my website or in the Google SWR spreadsheet. Please let me know if people are interested in that kind of information.

Also, one alpha source doesn’t rule out another. I still employ my options trading strategy, which adds about 4-5% annualized returns in my taxable account (about 1.6% spread over all accounts). Maybe combining the two would be the ultimate Sequence of Returns Risk killer!?

OK, we’re at 5,000 words now. If you read this far, kudos to you! Happy retirement, everyone!

Please leave your comments and suggestions below! Additionally, be sure to explore the other parts of the series; see here for a guide to the different parts so far!

Title Picture Credit: WordPress AI + ERN edits

Technical Appendix:

A few short technical notes that didn’t make it into the post

• Of course, I played around with many other momentum construction rules. For example, instead of the hierarchical rule gold -> equities -> bonds -> cash, simply allocate all unused weight to cash. And other rules. All methods I tried yielded very similar results, with Sharpe Ratios differing by no more than 0.03 from the preferred method I eventually used. I don’t include all those additional simulation results in the post because we’re already exceeding 5,000 words.
• I also simulated returns for a scenario in which we delay the momentum signal by one full month. That’s to address the concern that you would have traded at the market close every month, but waiting until you know the market close to plug that into your momentum signal, it would be too late to trade at that point. I found that even if I trade this momentum strategy with a 1-month delay, you would still have done okay. You face a deterioration in the Sharp by about 0.07 to 0.09. The drawdowns were still about the same. In practice, of course, you wouldn’t wait a full month. You might trade on the first trading day of the next month or even measure the momentum signals a few minutes before market close and get your trades executed that same day. But having worked in tactical asset allocation many years ago, I learned that it’s always a good exercise to see how quickly your sgianls deteriorate. This signal seems to be very robust to delays.

Like this:

Like Loading…

Related