Welcome to the follow-up to the follow-up post on the “Yield Illusion.” Again, here’s the context: a few weeks ago, I wrote a post (SWR Series Part 29) on why I don’t believe that chasing higher yields is necessarily a good hedge against Sequence of Return Risk. A very well-received post! It was picked up by CanIRetireYet.com as one of their Best of the Web in February, it was featured on RockstarFinance on Monday, and we had a great discussion in the comments section. So I wrote a follow-up post on Monday (SWR Part 30) and since that post was running way too long already, here’s some more material that got cut; some more thoughts on my asset class outlook, international vs. U.S. stocks, dividend vs. value stocks, and more. So let’s get rolling…
Some more thoughts on my asset class outlook
In the post two days ago, I certainly conceded that the failure of the Yield Shield doesn’t necessarily translate into failure during future bear markets. Past performance is no guarantee for future results! Because this is such an important issue and there is an important difference between the Yield Shield performance for fixed income vs. equities I like to stress once more how I conceptually think about this issue, see the diagram below. It’s for illustrative purposes only, this is nothing quantitative. It’s how I think about the prospects of the Yield Shield going forward:
- The equity recommendations of the Yield Shield are a crapshoot. Sometimes they work (dot com bust) sometimes they don’t (Great Recession). The high-dividend fund VYM (Vanguard) underperformed during 2008/9. But the SDY (State Street/SPDR) outperformed the broad index (more on that one below). I wouldn’t call this a solution to Sequence Risk. Certainly, the out-/under-performance is unrelated to the dividend yield! Note that within the dividend-focused ETFs, the SDY with the lower yield outperformed the VYM with the higher yield.
- The fixed income recommendations of the Yield Shield are just plain bad. The credit premium that gives you the higher yield will backfire in a bear market because this is essentially backdoor equity beta. Even the midpoint of the distribution is in the “make SoRR Worse” region. And Preferred Shares got hammered particularly badly in the GFC. But even if the next recession is not centered in the financial sector, they will very likely underperform the plain old safe government bonds. Remember, they are called Preferred Stocks, not Preferred Bonds!
Therefore, neither the equity nor the fixed income portion of the Yield Shield satisfies what I’m actually looking for in a Sequence Risk Shield! For that, I’d have to have a near guarantee of Sequence Risk protection, see below. A cash cushion or a glidepath may provide this, but don’t get your hopes up too high! That would only have a very small impact!
A few more words on Dividend Yield vs. Value vs. Quality
Quite intriguingly, a lot of the discussion in the comments section under the initial post centered on the equity dividend yield approach. As I said before, replacing a broad equity index with a U.S. high-dividend-yield equity ETF is probably one of the more innocent flaws of the Yield Shield method. Even I’ve played around with my own little dividend yield portfolio in my M1 account. But so far, it’s with play money only! Commenter Pat G. pointed out that there are other dividend strategies that look more promising. State Street (SPDR) has an ETF (SDY) that tracks only dividend payers that pay dividends consistently:
“The Index screens for companies that have consistently increased their dividend for at least 20 consecutive years, and weights the stocks by yield. Due to the index screen for 20 years of consecutively raising dividends, stocks included in the Index have both capital growth and dividend income characteristics, as opposed to stocks that are pure yield”
If we look at the return stats, the SDY outperformed the broad indexes (IVV, VTI) and certainly did better than the VYM. It did so by having the lowest VTI beta in the factor models I computed here and a relatively high bond beta (the highest among the non-REIT equity ETFs). It also has a strong Value bias (Fama-French HML factor). Even without the alpha, this would have done pretty well in 2007-09!
Side note: How strong was the value bias? Since the VTI beta was 0.8 and the HML beta is 0.4, we’re about halfway between an equity value portfolio and a neutral portfolio. A growth portfolio would have had a beta of -0.8. This shows, that there is a high but not a perfect correlation between dividend yield and value. Value is tied to the price-to-book ratio, not dividends!
But notice when you try to bring in an element of quality/value, you then have significantly lower dividend yields than the VYM used in the Yield Shield. SDY has a dividend yield of only 2.57%. Roughly half-way in between the IVV (about 2%) and 3.20% in the VYM!
So, just to be clear, I can imagine that consistent dividend payers are of interest from a Sequence Risk perspective. They are more likely to have stable profits, lower volatility, lower market beta, lower payout ratios, low debt ratios, better cash flow ratios, etc., so they may not suffer quite as badly in the next recession. A caveat though is that due to the popularity of ETFs and smart beta investing, some of this advantage will likely be arbitraged away over time! And I’d be a bit afraid of the “fallen angel” issue: What happens if a previously consistent dividend payer drops out of the dividend aristocrats? What if they first try to avoid cutting dividends, then drain their cash flow for years and eventually do cut dividends and get clobbered like GE?
Either way, I will not judge anyone for constructing a diversified quality-dividend portfolio! Maybe even replicate the constituents in the SDY in an M1 account and save the 0.35% expense ratio in that ETF! But someone too blind and too delusional in their yield-chasing effort might be exposed to the “dividend traps” like CenturyLink, which had one of the highest dividend yields in the S&P500, and got clobbered after announcing a 50% dividend cut. Incidentally, that happened a day after I published my initial post on the Yield Shield. You almost can’t make that up! 🙂 And, by the way, CenturyLink is in the VYM, but not in the SDY!
Some more thoughts on International Stocks
Are non-US investors doomed? Some commenters raised this issue. While it’s certainly true that international stocks underperformed I’m not saying that non-US investors should completely dump their domestic stocks. But there’s ample evidence that foreign investors have more to gain from international diversification than U.S. investors. Vanguard had a nice white paper making exactly this point, look at the chart on page 5. It goes back to the point I made in my 2017 post on diversification:
“When the U.S. economy struggles the whole world will feel the effects. In contrast, smaller countries can have recessions and market crashes without much of an impact on the rest of the world. So if you live outside the U.S., you’d greatly benefit from investing outside your own country!”
Some of the differential performance between U.S. and non-U.S. stocks also goes away when accounting for the FX moves (as discussed on Monday). So, for example, while Australian stocks performed quite poorly post-2008, measured in USD, they did a little bit better in local currency, and that’s what matters for Australian investors who live and consume there. But Australians certainly would have benefitted from holding U.S. stocks. That would have smoothed out the performance over the last two decades: lower returns during the early 2000s (when Aussie stocks went through the roof) and better returns post-2008 when Aussie stocks were struggling. Exactly what they find in the Vanguard White paper!
But, but, but, have you tried X or Y or Z?
OK, so people are throwing suggestions on how to salvage the Yield Shield at me faster than I can run simulations. Let me get one thing really clear here: The inventors of the Yield Shield had a ton of time and the benefit of hindsight to get this right. They didn’t get it right and I showed it. It’s not my job now to disprove every single variation of the Yield Shield that might work slightly better. Indeed, among the gazillion of different ways of tweaking the Yield Shield, I know for sure that some will work. Again, with the benefit of hindsight, maybe someone will finally tweak it so it works beautifully in 2007-2019. It doesn’t disprove my initial claim, which was “How Can a High-Dividend Portfolio Exacerbate Sequence Risk?” The title wasn’t “How every single High-Dividend Portfolio fails every single year in every single asset class.”
So, any additional comments of the type “Have you tried XYZ?” will be answered with this response: “No, I haven’t tried XYZ. Why don’t you try it and show me how it worked between October 2007 and now?” It’s not my job to contradict every single cockamamie Yield Shield variation now!
That said, here’s one idea that was floated before I even wrote the initial Yield Shield and that was again mentioned in the comments section on Monday. Why not invest 100% in a high-yielding equity fund, e.g., EWA, the Australian index ETF from iShares? Or the VEU (non-U.S. equity ETF)?
My response: It may certainly work going forward. But between 2007 and now it didn’t! Here’s the performance of three portfolios 1) IVV (S&P500), 2) EWA (Australia), 3) VEU (non-U.S.) between the May 2007 and Feb 2019 with 4% withdrawals (adjusted for CPI). The chart is generated with PortfolioVisualizer, see here for the permalink with all my parameters:
Both high-yield portfolios would have lost substantially. And the numbers are all in nominal terms so the 100% EWA portfolio would be down by about 45% in real terms if we deduct the cumulative 22% CPI inflation during this time. The VEU is then down by about 55% after CPI. And the low-yielding IVV? Would be back to the pre-GFC peak, even after inflation! Go figure!
And, to really beat this dead horse a little more, we can also do another little tweak. Live off your dividend income only but never sell any shares. This is what a commenter (“Torch Red”) suggested in Part 30:
“I am suggesting that retiring on a dividends-only strategy on a 100% stock dividend portfolio with initial yield of 4% eliminates sequence risk (as defined as selling shares when prices are low), establishes a clear SWR each year (set equal to dividend income), and eliminates timing the market each quarter (since shares are never sold). The downside to the strategy is that some diversification is lost and the retiree must deal with choppy income from year to year, though the long-term dividend dollar income will grow faster than inflation.”
I’m doing this for the Australia ETF (EWA) only. If someone wants to repeat the exercise for VEU, please go ahead. But this is a lot of manual work and given that the VEU with a constant 4% WR underperforms the EWA there isn’t much hope for this to work better. So, imagine on May 31, 2007, you’d bought one share of EWA and never sold it. Just live off the dividends. I pick the 5/31/2007 start date to be consistent with the other simulations in the earlier posts. But the results would look even worse if I had started closer to the market peak (October 2007). See the table below:
- Your income wasn’t enough to fund a 4% effective withdrawal rate ($1.12=$28.10×0.04) most of the years. But it was close. The dividends bounce around quite a bit but on average they were 17% below the target. But some years 40% below the target! Of course, if you have the flexibility to substantially cut your consumption in retirement, good for you. But notice that with this kind of flexibility, a low-yielding portfolio (either 60/40 or even a 100% S&P500 portfolio) would also make it fine through this Sequence Risk episode! And do much better than the EWA ETF!
- What’s even worse: that one ETF share you bought in 2007 never recovered. It’s still 23.7% below the initial portfolio value in nominal terms and a whopping 37.6% below the initial value in CPI-adjusted terms (which is really what matters!).
So, here’s the confusion: people incorrectly assume that dividends are never cut, they always keep up with inflation, and the price return not only recovers but also recovers to the previous peak plus inflation adjustments. None of those held true in the case of EWA. Or VEU, or USRT, or PFF!
But, but, but, the 2008/9 recession was so bad!!!
Some have pointed out that it’s “unfair” to expose the Yield Shield portfolio to the brutal reality of the worst recession is most folks’ lifetime. To which I’d reply: Well, isn’t that the idea? What use is it to propose a “solution” to Sequence Risk if it doesn’t work in the worst recession in most recent history? It’s like health insurance that only covers the common cold but has exclusions for heart attacks, stroke and cancer. If the Yield Shield works only in shallow, garden-variety bear markets, then what’s its use? In that situation, your trusted 60/40 or 70/30 or even an 80/20 portfolio wouldn’t be impacted too much by Sequence Risk either!
Yield Illusion vs. Yield Delusion
People here in the comments section here have been very kind and nice and civilized. Even if we occasionally disagree on certain issues. I’ve also tried to keep up with the attacks and insults directed at my work in the comments section over at that other blog and it got out of hands pretty quickly. It ain’t worth it! I’ve even gotten an anonymous hateful email over this! Then again, if it took three years of blogging for this to happen I must have been just too nice and agreeable before! 🙂
In any case, quite amazingly, despite the overwhelming evidence that the Yield Shield is really quite flawed and problematic there are still tons of folks who insist otherwise. They’ve turned the Yield Illusion into a Yield Delusion, and here are some examples:
- “But look at the YS portfolio post-2008; it recovered again” No it didn’t! I’ve seen some cheaply-put-together screenshots created with Portfolio Visualizer but the attempts to obfuscate the bad performance were pretty obvious: 1) the portfolio value wasn’t adjusted for inflation, 2) the withdrawals were 3.4%, not 4%, 3) the withdrawals were not adjusted for inflation, 4) the starting point of the simulation was not at the market peak, but in January 2008, leaving out the last few months with bad returns in 2007. So, when I show you that the Yield Shield portfolio is down $250,000 in real terms when applying the 4% Rule and others claim everything is A-OK, we know what explains the discrepancy: They are fudging the numbers.
- “REITs and International stocks outperformed the broad equity market in the early 2000s!” Yes, and this goes back to my clarification above: on the equity side, the Yield Shield is mostly a crapshoot. Sometimes it will outperform (1982-1987, 2000-2007) and sometimes it will underperform (the 1990s, GFC) and that has nothing to do with the dividend yield. That’s still a far cry from the promise of shielding/protecting you from recessions and bear markets.
- “The Yield Shield performed better than the passive portfolio when we start the simulation away from the market peak!” Exactly! That’s the definition of Sequence Risk. It can go in your favor as I showed in Part 14 of this series. If you pump up equity beta and equity markets keep going up, then the Yield Shield portfolio will outperform. That actually strengthens my claim that the Yield Shield portfolio exacerbates Sequence Risk (read again what the title of Part 29 was!!!): Both up and down!
- “I’m daring you and your audience to find a year X to 2019 where a Yield Chasing portfolio underperforms substantially (more than 5%) a 60/40 portfolio.” This is an actual comment from someone who doesn’t understand basic logic. I don’t have the burden of proof to do so. In addition to the bullet point above, recall that the title of the post was “How Can a High-Dividend Portfolio Exacerbate Sequence Risk?” and not “A High-Dividend Portfolio necessarily underperforms 60/40 by x% every single year”
- “The Yield Shield is only temporary and one can always move back to the passive 60/40!” Arghhh! Research on glidepaths, both here on my blog (Part 19 and Part 20 of this series) and the awesome work by Kitces and Kitces and Pfau points to the benefit of a glidepath from less equity exposure early on to increased equity beta. The Yield Shield does it the other way around: Start with equities “on steroids” and then back to a passive 60/40. Another reason why it’s a bad idea.
- “The Yield Shield would have worked better in combination with a cash cushion!” I know. I’ve done a whole lot of research on that issue (see Part 12, Part 24, Part 25 of this series). But that cash cushion would have also helped the passive 60/40 portfolio. Comparing apples to apples, the YS with a cash cushion also underperformed a 60/40 with cash cushion since 2007!
Whoah, this was a long post. Almost a manifesto (oh, no, let’s not go there)! Who knew this topic would get so controversial. Especially on the fixed income side, it seems pretty obvious that chasing yields to hedge(!) against the next recession is a bad idea. I showed this idea to some of my fixed income buddies and they got a good laugh out of it! On the equity side, at least you’re not handicapping your portfolio as with your bond portfolio. But crossing your fingers and hoping that the next bear market will look more like 2000-2003 and less like 2007-2009 isn’t a strategy for dealing with Sequence Risk.
So, I think I’ve pretty thoroughly debunked high-yield as a tool to alleviate Sequence Risk. The folks who had reached out to me and asked me to write about my thoughts on the Yield Shield certainly agreed with this. Probably most regular readers here on my blog agree. If there are still some folks out there who don’t agree I wish them good luck. I actually mean it! I hold Preferred Shares myself. I certainly don’t want them to hurt anytime soon. But at least I know about the risks!
Hope you enjoyed today’s post. Please leave your comments and suggestions below an make sure you check out the other parts of this series!
- Part 1: Introduction
- Part 2: Some more research on capital preservation vs. capital depletion
- Part 3: Safe withdrawal rates in different equity valuation regimes
- Part 4: The impact of Social Security benefits
- Part 5: Changing the Cost-of-Living Adjustment (COLA) assumptions
- Part 6: A case study: 2000-2016
- Part 7: A DIY withdrawal rate toolbox (via Google Sheets)
- Part 8: A Technical Appendix
- Part 9: Dynamic withdrawal rates (Guyton-Klinger)
- Part 10: Debunking Guyton-Klinger some more
- Part 11: Six criteria to grade dynamic withdrawal rules
- Part 12: Six reasons to be suspicious about the “Cash Cushion“
- Part 13: Dynamic Stock-Bond Allocation through Prime Harvesting
- Part 14: Sequence of Return Risk
- Part 15: More Thoughts on Sequence of Return Risk
- Part 16: Early Retirement in a low return environment (The Bogle scenario!)
- Part 17: Why we should call the 4% Rule the “4% Rule of Thumb”
- Part 18: Flexibility and the Mechanics of CAPE-Based Rules
- Part 19: Equity Glidepaths in Retirement
- Part 20: More thoughts on Equity Glidepaths
- Part 21: Mortgages and Early Retirement don’t mix!
- Part 22: Can the “Simple Math” make retirement more difficult?
- Part 23: Flexibility and Side Hustles!
- Part 24: Flexibility Myths vs. Reality
- Part 25: More Flexibility Myths
- Part 26: Ten things the “Makers” of the 4% Rule don’t want you to know
- Part 27: Why is Retirement Harder than Saving for Retirement?
- Part 28: An updated Google Sheet DIY Withdrawal Rate Toolbox
- Part 29: The Yield Illusion: How Can a High-Dividend Portfolio Exacerbate Sequence Risk?
- Part 30: The Yield Illusion Follow-Up
- Part 31: The Yield Illusion (or Delusion?): Another Follow-Up!
Picture Credit: pixabay.com