April 22, 2020 – Recently, I wrote a post endorsing the simple Bogleheads approach: invest in passive index ETFs. Everything else is just mumbo-jumbo, window-dressing and people not understanding the (mostly) efficient market nature of the stock market. In other words…
Simple (indexing) beats complicated active investing
Well, after unloading on some of the fancy complicated investing styles, I just like to point out the select few of them that indeed performed relatively well in 2020. At least better than the index. So, for the record, I’d also like to write about three examples where…
Complicated beats simple index investing
And most importantly, I’m not pulling some “Monday Morning Quarterback” nonsense telling you that if you could have sold your airline stocks in February and replaced them with stocks for video conferencing makers you could have done really well. Well, duh, very few people other than U.S. Senators had that kind of inside information back in February! Rather, I want to write about some of the deviations from simple indexing that were mentioned here on the blog in my posts and/or in the comments. Before the crisis!
Let’s take a look:
Let’s start with the one slightly less simple equity investing style I implement myself:
1: Equity Index Put Writing
I introduced the strategy in my option-writing series, see this article: “Passive income through option writing: Part 3” for the most recent update on exactly what I’m doing. The strategy involves selling downside insurance (put options) on the S&P 500 equity index and collecting a small but steady premium for that. It’s a great source of income in retirement, but just to be sure, it also involves the risk of some ugly losses once in a while when the market takes a sudden and sharp downturn.
How did we do in 2020? The actual put writing strategy is up by 2.6% year-to-date up to 4/21/2020. Not bad considering the crazy market moves and a loss of 14.8% YTD. The strategy lost a good chunk of money on February 24 but then rallied(!!!) in March with a +4% return, one of the best monthly results of the put writing strategy ever. I’ve had a 1.000 “betting average” i.e., I made the maximum premium on every single put option I sold since February 24. That includes the crazy days in March with a 12% daily drop. So, the strategy delivered exactly what I wrote about back in 2016:
“[T]he advantage of our strategy has been that if additional drawdowns occur after the initial event […], we actually make money. That’s because investors were in panic mode and drove the option premiums up so high that we sold puts at strike prices far out of the money: none of our short puts lost money even when the market dropped further in consequent weeks.”
Specifically, even at the height of the crisis, I was selling put options with strikes well below 2,000 points and since the market never fell enough I always earned the full premium
As with every derivatives/margin strategy you also have to decide what you do with your margin cash. I had really great success last year holdings most of that money in a) individual preferred shares (mostly financials, like Goldman Sachs, Wells Fargo, State Street, etc.), b) Muni Bond Closed-End Funds and c) Muni Bond mutual funds. Unfortunately, this year all three asset classes took a bit of a beating. They took my nice solid +2.6% Put Writing return and turned it into a -6.9% return in the portfolio overall. Well, you can’t win all the time! The Lord giveth, the Lord taketh. But it’s still better than the equity market overall (-14.8%). I’m still ahead of a 75/25 portfolio with corporate bonds (-10.7%) and even slightly ahead of a 75/25 with U.S. Treasurys (-8.5%). One drawback is that during April, when the S&P recovered quite substantially, the put writing only made around 0.5% and the fixed income portfolio is still not recovering.
So, it’s a bit ironic that I was caught in a bit of a downdraft, not because of the scary, risky leveraged derivates trading but because of the old-fashioned assets in ostensibly “safer” traditional vehicles like fixed income mutual funds and ETFs. It’s also ironic because I warned about the risks of squeezing out more yield.
But I made the conscious decision that I could afford to take a bit more risk with my margin cash because I felt that I’ve dialed in the risk model in the put writing strategy well enough. It also just so happened that in late 2018 I got a chance to pick up a lot of the Preferred Shares and leveraged Muni Bond Closed-End Funds at a great discounted price and I was sitting on a big pile of unrealized capital gains in early 2020 after the 2019 rally. Ideally, I should have liquidated them and shifted to safer bonds (e.g., 10Y Treasuries like in the IEF ETF) before the crash. I would have made a killing this year. But who knew that in February? And 20+% of the gains would have been taxed away, too.
So, long story short, this approach performed a bit better than a completely passive portfolio. Lower drawdowns, less volatility and a better return than a 100% equity fund. With more than 8 months left in the year, I’m hopeful that we’ll get above the zero line again.
Onto the next style:
2: Stocks+Bonds plus leverage, plus long-VIX-futures
OK, that’s a mouthful! A reader recently asked me about my opinion about his portfolio – a very simple 3 fund portfolio that would have done remarkably well during the most recent market meltdown:
- 47% UPRO (=3x leverage U.S. large-cap equities)
- 47% TMF (=3x leverage U.S. Treasurys with 20+ maturity)
- 6% VXX (long VIX futures)
This portfolio combines two styles that I’ve always found quite intriguing:
- Using the negative correlation between stocks and bonds, one could take a 50/50 portfolio of stocks+bonds and lever it up to the max. I’ve seen some Bogleheads posts where they recommend 50% UPRO and 50% TMF (some even going as high as 55% UPRO). It has a similar flavor to the post I wrote a long time ago in 2016: “Lower risk through leverage“, i.e., find the tangency point along the efficient frontier (likely closer to 40% stocks and 60% if you ask me) and then scale up the whole thing to a risk/return level you’re comfortable with.
- In October 2017, I published a post on why the short-VIX strategy as in the now infamous XIV ETF seems way too risky. Sure enough, the ETF blew up in February 2018 and lost 96% in one day nd then shut down. Since the short-VIX strategy had such unattractive risk-adjusted returns one could deduce that one should do the exact opposite: add a little bit of long-VIX as a diversifier into a highly-leveraged equity portfolio due to the reliable negative correlation with the S&P 500 (as I surmised in the comments below that post).
And it turns out, this combination of the two flavors worked quite well in 2020. I simulated the original 47/47/6 allocation but also the 50/50/0 version (as seen on discussed heavily on Bogleheads, i.e., without the long-VIX ETF) and a 40/54/6 allocation (slightly overweight the bonds vs. stocks because stocks have higher volatility). And then compare the three flavors to 100% equities invested in a passive fund, e.g., IVV.
In the table below, I display the simulated returns not just for 2020 YTD, but also for the 2019 blockbuster equity return year and the (partial) 2018 calendar year. I have to start in February because the VXX ETF didn’t start until then:
- The returns of the leveraged portfolios look pretty amazing this year: +11.1% for the 47/47/6. Without the diversification from the long-VIX you do a little worse, but you’re still up 0.3% for the year. And with a more cautious stock/bond allocation (40/54/6) you’d have made a killing: +21.3% during a bear market!
- This combination of funds would have also worked phenomenally well in 2019, but at the cost of slightly underperforming the S&P 500 in 2018.
- I also included the correlation matrix. Note the significant negative correlation between stocks and bonds (-0.36 UPRO vs. TMF) and the -0.93 (!!!) correlation between VXX and UPRO.
- The risk (measured as the annualized standard deviations) of the leveraged ETF strategy is substantial. That’s the #1 reason the strategy is probably not a good route for most investors. Even the 40/54/6 version still has a higher risk estimate than a 100% equity portfolio, but at least it’s in the same ballpark.
For full disclosure: this kind of strategy would not do very well in a “stagflation” environment when both stocks and bonds tank, think the 1970s and early 80s. But, knock on wood, the current crisis still looks like the stereotypical demand shock recession. If you don’t believe me, here’s a screenshot from April 20 when the crude oil futures turned negative!
3: Equity Index LEAPS
LEAPS stands for Long-Term Equity Anticipation Securities. And the LEAPS approach I found quite intriguing involves the exact opposite of the put writing strategy in #1. Very short-dated puts are often overpriced because myopic, overly risk-averse investors are scared about short-term fluctuations. But longer-dated options have comparatively low premiums. I never wrote a separate post about this but the topic came up in the discussions frequently, most recently after the Bear Market post last year (see here for the link to the comment).
The idea here is that since a call option premium is relatively low, one could buy a LEAPS-call with a strike at-the-money (i.e., near the current index value) and thus fully participate in all the upside potential of the S&P 500 index over the next, say, 2 years. But you wouldn’t lose your shirt if the index were to go down. You can never lose more than the premium no matter how far the index falls. Sweet! But, of course, you have to pay for this protection! 1) you pay the option premium and 2) you participate in the S&P500 price index only, i.e., you miss out on a roughly 2% annualized dividend yield.
So, let’s go through the following simulation exercise:
- On 12/31/2019, the S&P500 stood at 3,230.78 points. Let’s go to the next higher call option strike, 3,250. An investor with a $325,000 portfolio, buys a single SPX Call Option with a strike 3,250 and an expiration on 12/17/2021. The option has a 100x multiplier, so you have the upside potential of a $325,000 portfolio invested in the S&P 500.
- The cost of that option was about $27,300. I initially worked with an estimate of $30,000 but a reader was kind enough to look up the exact quote as of 12/31/2019 (from Bloomberg, I presume).
- The remaining $297,700 of the portfolio can go into some interest-bearing assets. I will look at four different alternatives:
- A money market account, paying 2% p.a. for January through March and 1.0% p.a. during April.
- Short-term Treasuries (1-3 years, e.g., the iShares ETF with ticker SHY)
- Intermediate-term Treasuries (7-10 years, e.g., the iShares ETF with ticker IEF)
- Long-term Treasuries (20+ years, e.g., the iShares ETF with ticker TLT)
Let’s see how this would have worked out in 2020. First of all, the Call option lost a lot of value. As of 4/21/2020, the value of the call option is down to $12,970. That sounds like a big loss, more than 50% of the premium we paid, but relative to the $325,000 portfolio it’s a loss of “only” 4.4%. In addition, your $297,700 margin cash would have stayed afloat and even gained a little during that period. It all depends on what you used for holding the margin cash.
- In all four scenarios, you’d have substantially been better off than with the plain 100% index fund.
- The more U.S. Treasury bond duration you added to the portfolio the better the outcome. Ideally, with a 20+-year Treasury fund you’d have beaten the index by about 30 percentage points. +15.9% instead of -15%.
The risk of this strategy is that if you had gotten too greedy and jacked up the yield of the margin cash portfolio with too much credit risk, going into corporate bonds or even high-yield bonds, you might have lost a big chunk of the margin cash.
Another drawback of the strategy: That low premium for LEAPS holds only during relatively calm periods. Currently, the premium for an at-the-money call is over 7%. So, one could argue that the time to implement this strategy when you’re at or close to the all-time-high when implied volatility is low and the premium for the call option is low. Once the house is on fire it’s too late to buy insurance! 🙂
Bonus: Some other variations
- One could combine methods 1+3: sell short-term out-of-the-money puts for income but also buy a long-dated protective put as downside protection.
- In method 3, if the 4.5% annualized premium seems too steep, one could also do a (vertical) spread-trade, i.e., sell a call with a higher strike to give up some of the upside in exchange for reducing the premium outlay.
- One could use the long-VIX futures strategy to hedge the downside of the put writing strategy.
- For large enough portfolios, say at least $300k, ideally $500k and more, one could easily implement method 2 without the high ETF expense ratios: simply use S&P 500, Treasury Bond and VIX futures at a small fraction of the cost.
Conclusion
What do all three flavors have in common? Everything is entirely based on index investing. No stock picking, no sector picking, no style picking, no dividend yield, small-cap, value, small-cap-value, tilt that might all be a hit or miss during a bear market (and probably mostly a “miss” during this current bear market as I pointed out recently). So, just like the Bogelheads and much of the FI community, I’m a fan of keeping it simple and working with a broad index, like the S&P 500.
But there are some active approaches that are still within the index universe. You simply do some finetuning on how you deploy that index. In one case you work with options to pick up a vol premium that you can’t capture with a bland old index fund alone. In one case you work with leverage to expand the efficient frontier. And in another case you do some simple market timing and hedge the market downside when the index is goes from one all-time-high to another and the premium for an at-the-money LEAPS call is low.
Hope you enjoyed today’s post! Looking forward to your comments and suggestions below!
Title Picture: pixabay.com
Hi Ern, nice article. The spx 3250 leap was 273 on 12/31/19, fyi. Your estimate was pretty close.
Thanks, John! I will update the calculations!
Aren’t all the leveraged products calculated on a daily basis.
So UPRO in a volatile environment would need to be calculated daily to show outperformance and would cause losses. What am I missing?
These are monthly returns of the ETFs from YahooFinance, compounding the daily returns.
You’re right, all the exotic ETFs are rebalanced every day, so you will never find that the monthly return is exactly 3x the monthly return of the underlying.
Thanks BigErn on the excellent article,
I was looking forward to hear your opinion about including long Volatility as part of the portfolio as I have noticed a published paper on this topic by Chris Cole of Artemis Capital :
https://www.macrovoices.com/guest-content/list-guest-publications/3502-the-allegory-of-the-hawk-and-serpent/file
Thanks again.
They warn that short-vol will always fail. Strange! I think the approach has to be more nuanced. Normally, getting wiped out is a result of a poor risk model (see XIV ETF mentioned in their article too).
My personal recommendation:
Do short-vol with very short-dated contracts (put writing 3x a week) but only if you know what you’re doing. 🙂
Do long-vol over long horizons (1m+ with VIX futures, 2Y out with protective puts or long-call).
“I was sitting on a big pile of unrealized capital gains in early 2020 after the 2019 rally. Ideally, I should have liquidated them and shifted to safer bonds (e.g., 10Y Treasuries like in the IEF ETF) before the crash. I would have made a killing this year. But who knew that in February? And 20+% of the gains would have been taxed away, too.”
I am facing this situation right now. Despite the market losses this year, by far my largest equity holding (in my former employer) has rallied to all time highs. I could liquidate and further diversify into S&P 500, but I’d have to pay 23.8% federal tax and a 10+% state tax. Very tough decision. This situation has presented itself numerous times over the past 15+ years, and each time would have been the wrong move (I did diversify over that time period and lost out on massive gains).
I would keep it, unless it’s a very high percentage of your NW. By selling now (vs. in retirement with low LTCG tax) you’re giving up a guaranteed 23.8%-33.8% return (assuming you’ll pay 0%-10% LTCG tax in retirement) in exchange for less risk via diversification.
Agree. 🙂
Yeah, it’s a “good problem” to have. With tax rates that high, it’s probably best to keep everything as is for now. 🙂
Hi Ern, nice article. I am trading the 47/47/6 portfolio, weekly rebalance, so far it has done well during this crisis. I am not so confident on what can TMF do over the long run , but for now I am just sticking to the rules.
I wanted to ask you about the short term put selling strategy. I think you choose the strikes by picking an annualized return near 5%, that would be independent from the delta of the option. In that case I don’t understand how can you get a 4% return in 1 month, even if all trades expire worthless. Or is it you choose the option based on a 0,10 delta?
On the other hand, do you think that the permanent portfolio can be a good base to sell this short term puts on top of it ?
Many thanks for sharing your work
Thanks! I believe you sent me the email and pointed this out to me right?
Agree with the TMF: Not much return potential. It will get hammered if things normalize again and yields rise again. But: then your UPRO kicks in and you make tons of money from that.
About the put writing: The 5% annual yield and 5-Delta are rough guides. When vol is high your 5-Delta put pays much more premium than 5% p.a. The last few weeks I’ve been writing puts with a 2-Delta.
I’m not a huge fan of the permanent portfolio on a stand-alone basis (https://earlyretirementnow.com/2020/01/08/gold-hedge-against-sequence-risk-swr-series-part-34/), but it might be stable enough to keep as the margin cash for the put writing. Yeah, that might work! 🙂
Hi, yes I wrote the email.
I am starting to trade the short term dated puts in SPY for learning purposes. For example as of today, a 0,05 delta for Friday pays 17,5% annualized , and a 5% annualized is roughly a 0.02 delta. If you always stick with a 5%, as volatility rises you are able to lower the deltas, given your experience doing this, sticking always to the 5% annualized return does makes sense over the long run ?
Thanks !
That’s what I’ve been doing. Has served me well since 2011. The strikes go so far OTM after the initial shock that you’ll likely make money after the initial loss. As in Q1 2020.
Wondering if you care to comment on the risk in the margin cash bond portfolio WRT defaults? Do u think the shutdowns will drive a bunch of corporate, municipal defaults and if so how do you think that will impact you?
That’s possible. Even though, history teaches us that Washington cannot afford to let states go bankrupt. This is being floated as a bargaining chip.
Corporates can and will go under, though. I suspect that the comanies that I have in my portfolio, GS, WFC, STT, MS, USB, C, will survive, so the preferreds will pay as promised.
I’d like to see your opinion on the dual momentum strategy discussed in Gary Antonacci’s book (if you are familiar). His strategy is advertised to limit the downside of drawdowns by shifting to T-bills and using relative and absolute momentum to purchase the index based on 12-month momentum.
It sounds promising, but I always am skeptical of “too good to be true” portfolio strategies. Some of my concerns are around implimentation costs and “whipsaw” market environments.
I haven’t read the book. I wrote about a simple momentum rule here:
https://earlyretirementnow.com/2018/04/25/market-timing-and-risk-management-part-2-momentum/
The tricky part is finding the right rule and sticking with it. In this current bear market, the SPX dropped below the 200d line on Feb 27. Then went above again on 3/2 and then dropped again on 3/3. So you were whipsawed just a little bit but mostly it worked.
The problem: now you’re out and the momentum signal will not go positive again until way later. You may miss the entry point.
So, generally I feel exactly like you: intrigued but not really comfortable acting on it due to the fear of the whipsaw.
My impression of why you use momentum is not to try to avoid all of a move down or capture all of a move up, but rather to avoid the worst drawdowns and from time to time catch an updraft. If the outcome of a sharp drop and rebound is that I am exactly as wealthy as if I stuck with b&h, then it worked.
Ive been using the method described in the book DIY financial management by the founders of Alpha Architect; they use a 50/50 time-series-momentum/SMA approach for risk management. It worked OK; got me out of most equity positions at the end of Feb, with 50% left in some equity etfs. Then as things continued to drop I short-circuited the rules and sold the rest mid-march, and then got a little FOMO and short circuited the rules again to get a little back in early April. So basically I’m a poster child for ERN’s statement that it’s tough to stick with. Financially, it hasn’t hurt me relative to b&h, but I have learned a lot about my own emotional reactions to market moves.
Last year it DID hurt me as I stayed out of the market when the rules said to be in! So what I’ve learned is that if I keep a portion in b&h I don’t get FOMO, and if I keep a portion in a momentum strategy I don’t panic sell. But then
…but then I’ve complicated portfolio management for myself.
Interesting! I agree that sticking to a momentum strategy could be much easier said-than-done as you’ve suggested. Perhaps if the rules could be completely automated we could get around some of the human elements. I’d really like to see someone who has successfully carried this out using real data compared to simply 80/20 B&H. Strategies like this sound good in theory, but my concern is once you factor in real life frictions, the benefits may be greatly reduced, thus going back to the KISS B&H approach.
I also don’t think I’d ever put my entire portfolio in any complex strategy. I like the concept of ERN’s put-selling strategy, but then I find it strange that ETFs that attempt to implement a similar strategy appear to greatly underperform simple indexing.
I also am pretty lazy, so I would have to be really convinced the benefits of any active strategy would add value or the magical alpha, and ideally convinced the strategy would be persistent. While momentum has a good track record, the explanation for persistence moving forward is a little weak IMO.
If you want to dig into the learning a little, I’ve found some value in Meb Faber’s blog. He has a free book and a bunch of white papers that are quick reads on quantitative value / momentum strategies. He also runs Cambria Funds, which have some ETFs that attempt to automate these strategies across a global asset allocation. You can look at TRTY, for example, and see that yeah, it underperformed the heck out of SPY; but it also barely even budges when SPY tanks. Which sort of makes sense when you realize that on the one hand, something like TRTY is basically the global asset with value, momentum and trend-following strategies all mixed together; the list of holdings reads like almost everything in the world that can be financialized in one ticker, and it doesn’t really seem to grow all that fast. And on the other hand, SPY is essentially a 20% allocation to 5 US tech companies (MSFT + AAPL + AMZN + GOOGL + FB) and an 80% allocation to a bunch of other stuff. Those 5 names have been great to anyone holding SPY (usually).
Alpha Architect is also a good resource; they published a book call DIY Financial Advisor that I found pretty easy to wrap my non-phd head around. Their approach is more concentrated, less kitchen-sink-esque. And they have ETFs – I mean, who doesn’t these days?
TRTY: Yeah, half the standard deviation. But the performance stinks. It makes me think that a lot of the “trendy” styles (pun intended) might be overfitted to look really good in backtests. But they stop working in real-time.
Oh, wow, thanks for sharing that! This is extremely useful! I wonder if some of the people who got out at the bottom in mid-March are still out today. So, people destroyed the momentum alpha from the first sell signal.
As it happens, we are still most of the way out, but were also not all the way in to begin with; we had a large cash position b/c we were planning a big home remodel (demo’d most of the interior of our house right before stay-at-home went into effect which is a whole story on it’s own). So call our equity position on Feb 1 52%, with 25% in cash and the balance in bonds, REITs and commodities. We’re now probably about 15-20% equities, though TBH I don’t have a precise breakdown at the moment.
ALL of that is to say, from Feb 1 to right now, we’re down 9% on our financial holdings. I’d say not too bad, could be better. If it was straight 100% [ERN: corrected the typo :)] B&H the SPY we’d be down about 14%, if it was 60/40 we’d be in a better position. All in all, the larger regret, for me, is missing most of the 2019 gain rather than taking the 9% loss from the highs. So much depends on what the next few months look like (when do the major indexes cross the 200-day, whether there is a “retest” etc. The market has a way of making you feel like a dummy no matter what you do.
* Straight 100% B&H the SPY, not 10%.
Yeah, that’s what I assumed! 🙂
Good point. If you look at your combined 2019/20 return even a 100% equity portfolio looks much less scary. Same with 2017/18, by the way! 🙂
Hi ERN, great article, as usuall on your blog. In the past I also thought I could use similar approach #2 but I don’t like this environment when bonds are pretty high with very low yields and zero interest rates (I also don’t like these leveraged ETFs as well), as soon as the dust settles bonds will be a drag in long term in my view. I am not sure how this defense approach (over 50% in bonds leveraged) will do in recovery which will happen sometimes…
In case of such a high negative correlation between UPRO and VXX, there seems to be no reason using VXX and instead decrease share of UPRO, which should do the same as using VXX.
Yeah, I hear ya. In absolute terms, bond yields were low back then. Nobody could have known how this might have worked out.
Oh, during the recovery, it works out exactly like this: you lose with your leveraged bond ETF but you more than make up for it with your leveraged equity ETF.
Disagree with the view on VXX. There is a case to be made for VXX. The cost of VXX is less than the opportunity cost of reducing your equity beta (recall: XIV had a lower Sharpe Ratio than the equity index)
I have been trading N2 for a few months already. The difference I see with VXX is the magnitude of the move, although correlation is almost -1, if stocks are hit very hard, a small allocation in vxx will prevent a big part of the losses.
In terms of TMF, I have the same fear given the environment, but hopefully it helps to smooth a bit the equity during the recovery, given the hole strategy is quite leveraged. I have also used TYD as a “less risky” leveraged bond option. This one shouldn’t suffer as much as TMF if bonds came down.
Something like NTSX seems similar to #2 here but dialed down in some ways. I was tempted to just put a chunk of change there and forget about it for a while, since the frequent rebalancing / attention cost for these was a little discouraging for me. Maybe I’m just too lazy?
Yeah, that’s more in line with my earlier post “lower risk through leverage”
Also: the 0.20% ER is pretty low. Not as low as DIY with a large account ($500k+) but good enough for small accounts.
Hi Karsten,
I have put your very safe option writing strategy to work for the last few weeks and has been working out very well and only about $140k at risk (correction – very little or no risk- :-)). I am still in pretty much testing phase with extreme conservatism, just so I have an opportunity to learn without risking too much.
Later on I will develop proper parameters and my own standard deviations, may be I will throw it on efficient frontier just to see how that fits (more or less for fun).
Thanks for your insights.
Karl
You bet! Best of luck.
Don’t put all your money into this in the beginning, though! 🙂
Excellent write up. Have you thought about how these different approaches could best offset sequence of return risk at early retirement?
Haven’t simulated that yet. It would be hard to get long-run data on put writing and VXX.
Great ideas. I appreciate all these posts on alternatives to passive index investing.
For the “stocks and bonds + leverage” plan, what about buying protective puts instead? I realize that in general buying puts won’t make money- just like buying insurance isn’t expected to make you money. But it would offset the worst risks of the leverage in a sudden crash, while the higher expected returns of the leverage pay for the puts.
Depends on the time to expiration. put buying is usually too expensive, except for the 2+-year LEAPS.
“one could easily implement method 2 without the high ETF expense ratios: simply use S&P 500, Treasury Bond and VIX futures at a small fraction of the cost.”
I’ve been doing something similar, but without the VIX futures (thanks for that idea).
One catch is that brokers make you keep a certain deposit on the futures in cash, and it seems they can change that amount whenever they feel like. Mine (interactive brokers) hiked the amount about four-fold during the crash. It’s pretty scary see your minimum balance suddenly quadruple while the stock market is also crashing. That’s made me pretty scared of using this sort of leverage going forward- just too much risk of a margin call.
Yeah, I’m using IB myself. Always keep much more margin cash than the minimum required!
Keeping it simple, large cap growth (VUG or SCHG) has out performed the total market as well the most few years. No energy or small caps dragging it down.
Yeah, essentially the opposite of Merriman’s portfolio. But you gotta admit: it wasn’t clear that this approach was bound to work during the upcoming recession. In January 2020, I would have thought tech-bubble-burst is one of the possible catalysts for the next recessino.
Hi, ERN. Regarding strategy 3, why choose at-the-money calls instead of deep in-the-money calls? Wouldn’t it be more advantageous to buy deep in-the-money to have more delta and less time decay?
It’s a different risk profile – the ITM calls provide less downside protection / risk reduction. Just playing with some data, if you had the 3250 call, you would be down 3.9% from the start of the year based on the notional value of the SPX at the start of the year. For a 2800 call (more in the money when bought), you would be down 6.7%. However if you held them to expiration, the 3250 call would obviously make less money if the market rose. I imagine the play might be to only hold it for 1 year before rolling to the following year call to try to keep losses from theta low.
Good answer! couldn’t have answered it better! 🙂
John had a good reply! 🙂
Oh man! Another juicy ERN article and about investment methods. This self isolation has brought about one good thing in life. More Articles. Thank you for this. I’m already dong #1. I’m interested in #2 for any future investments and I’ll keep #3 in the bank for when there’s low vol again.
Thanks, Bob! Stay healthy and good luck! 🙂
Hi Ern,
Instead of using VXX or VIX futures, how about buy a VIX call every month? CBOE has an article on their VXTH index (http://www.cboe.com/vxth ) based on holding the S&P500 and then buying VIX calls to hedge. It seems to be that their strategy can easily modified to hold the S&P500 with some leverage and then buy VIX calls. This seems similar to the second strategy.
Good suggestion! Yes, VIX calls would be another option. I haven’t done any analysis on what’s more efficient. Both strategies will be a drag on the performance if nothing bad happens, but then pay off royally when the crap hits the fan!
I really enjoyed listening to you on the recent ChooseFI podcast, and have been reading through many of your articles. Thank you for this wealth of information!
I was wondering if you heard the recent ChooseFI podcast #193 entitled “The Role of Bonds in a Portfolio with Frank Vasquez.” Frank made a very convincing case for using a bond fund in the portfolio that has a good negative correlation with equities. He specifically recommended the ETF with the ticker: “TLT” (or a similar ETF at Vanguard (ticker: “EDV”).
I read a previous article of yours in which you made a great argument against adding bonds to a portfolio, but I’m very curious to know your thoughts regarding that podcast’s advice. I would also love to know your opinion about using TLT or EDV for the bond portion of a stock/bond portfolio. Thank you!
I don’t like bonds in the accumulation phase and have written extensively on that.
In retirement, you certainly do need a diversifying asset like bonds. I use IEF (7-10 years) in my simulations, but TLT (20+ years) has even longer duration and thus even better diversification benefits.
Thank you for your reply!
I read with great interest your articles entitled “Have bonds lost their diversification potential?” and “When bonds are riskier than stocks.” I noticed that your analyses used bonds with much shorter duration than those contained in TLT and EDV. Would that potentially change the conclusions reached in your bond article analyses? Also, for someone in retirement, would you recommend using TLT and/or EDV for the bond portion of their portfolio? EDV has bonds of even longer duration than TLT.
The difference is that bonds never made much sense to me during the accumulation phase. You might as well go with 100% equities while accumulating.
When retired you need a diversifying asset.
I do the SWR simulations with IEF, but TLT or even EDV have even longer durations so they work potentially even better. I simply had the long return series for 10Y bonds.
Thanks!
I’m surprised and impressed that the put writing strategy continued to work tolerably in this situation.
I had questioned why someone who clearly wants a very secure (~ low WR), high income retirement with uncertain ability to make up losses by going back to work would adopt a strategy with such a lot of downside risk.
I wonder if the market’s overvaluing avoidance of downside risk will survive the next few years if this strategy becomes better known.
Either way, I am seriously considering using the strategy in retirement.
The strategy is pretty well-known, both in academic circles and among practitioners. Hedge funds have been doing variations of this for decades. The extreme negative skewness keeps the majority of investors (personal and institutional) away from this.
Then again, investors like this drip style return profile (which is probably why many fall prey to the overleveraged versions and don’t do enough DD) vs. Michael Burry’s long put approach… investors were pissed even after he made them rich.
Again: It depends on what kind of investor. Large investors like pension funds and endowments hate the negative skewness.
Hi Ern,
As usual very informative post and thorough analysis. I have a few questions/comments.
Use of VIX futures, seem to be only useful to offset volatility. In the longer run expected return from it should be zero.
Is that the right way to look at it.
On your comment that the second strategy can be built on S&P 500, Bond and VIX futures.
I am little worried here on two counts. In the very long run, without getting interest, how can a bond future have an expected return other than zero. Less of a concern that we will get no dividend from S&P 500 futures since the value of the future will have long-term positive expected return.
So is there a better way to do this without forgoing interest from bond (And if possible dividend from the stock index)
-Naren
“And in another case you do some simple market timing and hedge the market downside when the index is goes from one all-time-high to another and the premium for an at-the-money LEAPS call is low.”
You mean puts? 😉
Either one works. You could do the long-call and keep the margin cash in fixed income (as I showed). Or hold your portfolio in equities and buy a protective put. The two are equivalent, thanks to put-call-parity.
But the long call is more margin efficient. 🙂
Karsten, you mentioned that method 2 won’t work in stagflation and that the standard deviation is too high for most investors.
Does that mean you wouldn’t recommend it or see it as a good long-term strategy for those in accumulation phase or retirement phase?
Also if you hedged the put writing strategy, would 6% also be the amount you would allocate to the long VIX position?
I think method 2 would still be OK in the accumulation phase. If it goes against you, so be it. Just save a little bit longer. Method 2 would be a disaster for retirees if we were to have a stagflation environment.
The long-VIX position, as calibrated in the VXX, loses about 50% of its value when the market does well. So, 6% VXX would be a drag of 3% on the portfolio in a blockbuster year like 2019. But that was also the year when I made 90+% of my premiums. So, yeah, somewhere around 4-6% VXX might not be a bad hedge against a bad market.
Would method 2 then be a way to get to the retirement number faster, with the risk being stagflation, making accumulation slower. I guess what I’m trying to say is, is this a superior method than just plain index ETFs for accumulators like me?
Since the put writing accounts are larger anyway, would the cheapest/best way to implement the hedge would be to go long the VIX futures?
I’m also guessing we won’t need to hedge until the VIX goes back closer to the single digits?
1: I like method 2 for people who have very little to invest because they’re just starting their FIRE path.
2: Yes, that would be ideal. Lower expenses.
3: Absolutely! The long VIX approach is best done when the VIX is in the low-to-mid-teens again. Buy insurance when nobody worries about the risk! 🙂
1. Without you having to go into the math of it, is there a point where you’d recommend the plain index over method 2? Half way to FIRE? I guess the other thing is it would be hard to unwind this position due to baked in capital gains taxes so one would have to hold this almost indefinitely. Perhaps consuming this portfolio first during retirement.
Not a hard formula. But I’ve done some simulations on glidepaths pre-retirement where it would have been optimal to keep leverage until you reach about the half-way point in assets.
I’d find method 2 a bit too risky in retirement. Of course, in hindsight it did very well. But that’s mostly due to the leveraged bond portion. Maybe the long-VIX isn’t such a bad idea in retirement as a hedge. 🙂
I hope I’m not bugging you by asking so many questions but hopefully just one more!
What would be the best way to stop implementing method 2 at the halfway mark?
Because ideally, you’d be sitting on some hefty gains which would trigger taxes if we then switch this portfolio to a simple index one. Would we just stop contributing to this portfolio but let it ride and them try and consume it first in retirement?
If people have a high enough net worth and implement this through futures on IB, it’s pretty easy: Simply scale down your leverage over time by rolling into a smaller number of futures contracts (or simply go into the index ETFs over time).
The tax situation depends on your home country. You can always target a leverage of 1/(1-taxrate) as I proposed here:
https://earlyretirementnow.com/2016/06/07/synthetic-roth-ira/
So, another reason to implement this in your own futures account: more control over the leverage.
What about unwinding a UPRO/TMF/VXX portfolio?
I would also like to implement this through futures but I think the hard thing is most of my investment are already in index and in put writing. So this new method would require a drip of contributions over time. I would be able to contribute regularly to a UPRO/TMF/VXX ETF portfolio but it would be harder to have the $300-500K portfolio to implement with futures straight away.
Yeah, unwinding the UPRO/TMF/VXX means selling the assets. Unless you do this slowly and simply invest the new capital in something less risky and slowly phase out the risky stuff over time.
This has been working for me:
NTSX 60%
PHYS 25%
TMF 10%
TVIX 5%
Nice! Looks like a similar flavor as method 2 but with the addition of Gold. No doubt this worked well recently,
Long-term, I’d consider the 25% gold a bit too high, though. 🙂
Sorry if this has been asked in answered, but in tour simulations of UPRO/TMF/VXX, how often did you re-balance? IT seems given the volatility of all involved, it could get pretty out of whack on a day-to-day or week-to-week basis. Have you seen this paper on “rebalancing luck”?:
https://blog.thinknewfound.com/2019/11/the-dumb-timing-luck-of-smart-beta/
I would think this would be more important than average in this specification, no?
Coincidentally, I was about to ask this question today!
Can I add some additional questions then? How would you implement long VIX futures in method 1? Would you sell the futures every time the put writing experiences a loss and the VIX spikes? Then don’t go long VIX until it settles to < or = to mid teens again? Hope that question makes sense.
With the caveat that I’ve not implemented this myself, I’d do it like you proposed. Go long-VIX when times are calm.
The question is, what to do when there’s a spike in Vol. When do you stop? You might have stopped too early and missed the 80+ VIX. The devil is in the details! 🙂
That’s exactly my fear. I’m not smart enough to figure it out unless maybe by trial and error since I’m not a math person.
Will you post or comment about the implementation once you figure it out?
It’s not high on my priority list right now. 🙂
Monthly rebalancing.
If you have strongly trending markets, you’d get very out-of-whack weights even intra-month. But it’s the only clean return data I have available. 🙂
Yeah, nice link! Very interesting!
I assume the data is monthly so you might now be able to comment but are other rebalancing frequencies better than monthly? I guess the less frequent, the less trading fees introduced. What about quarterly as many in the Bogleheads thread seem to be implementing?
It’s a hit or miss. In trending markets you’re better off rebalancing less frequently. In choppy, mean-reverting markets you’re better off rebalancing more frequently.
Karsten, can you help me see if this post on Bogleheads makes sense regarding method 2?
https://www.bogleheads.org/forum/viewtopic.php?p=5304169#p5304169
The way this person is implementing treasury futures is targeting duration rather than a leverage. My amateur thought is that to gain 3 x leverage treasuries, one would buy a futures contract while holding 1/3 of the notional dollar amount as margin.
Is this poster correct in targeting duration instead and getting more than 3x notional leverage? I admit it’s a bit over my head.
Duration is more “accurate” because 3x leverage with 30-yeay bonds has a very different risk profile than 3x leverage of 2-year bonds.
So, calculating the duration of your bond portfolio (whether physical, ETF, leveraged ETF or futures) is the proper thing to do.
I see. How do I work out how to achieve the same 3x leverage TMF would get me if using futures instead?
I went back and read it again, the poster in Bogleheads is using one 2 year, one 10 year, and one 30 year futures contract to gain 3x leverage on $70K.
Looks like I have a lot of reading to do.
Btw, I am watching your youtube interview with Personal Finance Tips. I can’t get enough Big ERN content!
Haha, yes, that was a fun interview. Not a very well-known podcaster/blogger but I liked his interview style and he asked all the right questions! 🙂
TMF is also using futures (to my knowledge). So, to “replicate” the TMF one would simply replicate its holdings, which should be posted somewhere.
I’m glad you’re still giving small content creators your time haha
It’s good for me because I need all the Big ERN content I can get while in isolation.
I’m having at a looking at the TMF fact sheet, prospectus, and the daily holdings and I can’t seem to find the info. The daily holdings seem to show they use a lot of swaps rather than futures if I am reading it properly.
Is my simple notion that if I hold 1/3 of the notional dollar amount of an 30 year treasury future, that isn’t the same as the 3x leverage we want in the method 2 portfolio?
I can see that duration of TMF is about 57 years so we would want 3x leverage of a 19 year duration bond. I can see that TLT is about 19 years duration. Is a 30 year treasury bond 18 years duration or is it more difficult in calculating this?
Ah, OK, if hey use Swaps, that’s not something you can replicate as a home investor. That’s OTC between the big guys.
So, for the 3x leverage, if the long Treasury Future has a quote of 120, you’d need $40,000 principal and hold one such contract. That would be 3x leverage of the 30Y bond.
If a 10Y bond has a duration of about 8 (8 times 7 is about 56, close enough to 57), then you’d need around 7x leverage with the intermediate bond future. So, if that future has a value of 110, you’d need roughly 110,000/7=$15,700 of principal per long 10Y future.
Thanks again for working through this with me. This is the equivalent of a primary school teacher going over math to a child. So thank you for your patience.
I think I am almost there with my understanding on how to implement the bond portion through futures now but I have a few more questions.
1. TMF is roughly 3xTLT. Is the long treasury future equivalent to 1 TLT or is it more complicated than that?
2. Is the relationship of 7x leverage with the intermediate bond future always constant in giving us the same duration as TMF or can this change?
3. How do you find or calculate what the duration of a 10Y or any bond? Is a 10Y bond always about 8 years in duration?
Thanks for always taking your time to answer my questions Karsten 🙂
As far as I know TLT is indeed holding 99+% in long T Bonds, 20+ years, I believe: https://www.ishares.com/us/products/239454/TLT?cid=ppc:ishares_us:google:tickers-priorities&gclid=CjwKCAjwxev3BRBBEiwAiB_PWFJGy0tpVUfJbzYeJkYQv-p7gLYFXksIyG4cxzabk6cW3fzeUVaRoxoCmc4QAvD_BwE&gclsrc=aw.ds
Duration is a moving target. You can look up the TLT duration on the same page.
The 10Y duration has to be calculated for that specific bond. It’s probably somewhere around 8-9, but it’s changing over time depending on the specific bond that’s the cheapest to deliver for the 10Y Treasury Future. Not so easy to find precise current estimates unless you’re an investing pro with a Bloomberg terminal, I’d suspect. But 8 to 9 is probably a good estimate right now. 🙂
I see so if TLT is just 99+% long T bonds and TMF is 3x leverage TLT, then 3x leverage 30 year long bond futures (UB) pretty much replicated TMF?
That should be so, yes.
Am I interpreting right that you would prefer the 40/54/6 split for UPRO/TMF/VXX in method 2 rather than 47/47/6?
Does it matter that the current bond yields are so low?
No strong feelings. Depending on how you construct a portfolio you’ll get different results. If you set the weight proportional to 1/vol then you might need a slightly larger weight on TMF than UPRO.
But I’m sure you can come up with some other procedure to justify 47/47 considering the low bond yields.
I also have no strong feelings either. I think I’ll have to go back and read your analysis again and decide what kind of risk profile I’d like.
i am very interested in the tmf upro combination. unfortunately i can’t buy these products in germany. my broker said it’s because of the withholding tax. at least the upro i can still easily reproduce via warrants, possibly also via futures. at the tmf it gets harder. here i can’t even find warrants.
Maybe you have some tips for your followers from your german homeland ?
If you can trade futures, the TMF should work the same way.
Great post! Index investing all the way.
https://planonfire.com
Cool! We’re on the same page, Jacob!
Nice blog!
Vey interesting article !
The blog mentioned below will surely help readers and give a broader view on global equity market. Please share your views on the same.
https://fundamentalmarket.wordpress.com/2020/07/13/indian-equity-market-near-its-four-month-high/
Thanks! I don’t really dabble in the Indian market at all. But I hope people will find this interesting! 🙂
Hi Karsten,
I’ve been paper trading two versions of method 2.
1. Futures – ES, UB, VIX
2. ETFs – UPRO, TMF, VIXY
Both in 40/54/6 ratio.
I’m getting some strange results so I’d like to get your take on it. Firstly, I’m finding a few hundred dollar difference between the two paper $500K portfolios. This isn’t too bad since I put it down to less granularity in asset ratio for futures. I’m sure this can be fine tuned if I was serious about calculating it.
The strangest thing I’ve found though is where the gains and losses are coming from.
In portfolio one, UPRO has provided all the gains which helped put the portfolio in positive territory even if TMF had big losses.
In portfolio two, it’s the reverse. Gains in UB is making up for losses in ES.
Granted this is very casual experiment with little rigor, and I’m just basing numbers of Interactive Broker’s numbers.
With smaller portfolio sizes, we would probably have to go with Micro ES, TMF, and VIXY to replicate this portfolio so it would suck to get lossess in most MES and TMF if the above phenomenon happens.
Any thoughts on this?
Hard to tell without seeing the results.
It’s possible that there are large differences because UPRO, TMF have to rebalance every single day and you get some return differences from that. Momentum alpha when returns are drifting. And whipsaw losses when you have choppy markets.
I completely understand that you’re having to just make general comments especially since it’s just based on my non-rigorous paper trading.
If your postulation is correct, does that mean it is ill-advised to mix and match ETFs and futures for method 2?
In my paper trading experiment, I think I would be at a loss if I had used ES as my equity exposure, and TMF as my bond exposure.
Or at least less gains since the market is at all time highs agains.
Again, can’t comment on sims I haven’t seen. Maybe the ES future trading would be too granular in a small portfolio.
But I also think that the high fees in the TMF and UPRO ETFs could undi a lot of their advantage. Hard to tell…
When using ES, would you recommend allowing the contract to expire (cash settlement) and buy the next contract, OR would you roll the position before expiry?
There is a feature in IB to roll the contracts, i.e., a spread trade to sell the near contract and buy the next expiration.
When would you roll a future? I’ve read people try to keep an eye on volume? If the markets are efficient, then would it matter if one rolls the contract close to the last minute?
There is a window of about 1-2 weeks before the front contract expiration when both contracts are very liquid. I would not wait until the last minute.
Would method 3 be a good idea now that the VIX is back to lower levels?
Would you still recommend ATM strikes? I understand from John’s answer in the comments that ATM strikes would offer the most downside protection but the breakeven is higher.
What do you think about a Poor Man’s Covered Call strategy, buying a deep ITM strike (70delta) and then selling short DTE calls? The breakeven from a deep ITM call would be less as well.
Another thing aspect I’m worried about is that TLT is expected to go down if rates are ticking back up, would method 3 be a losing proposition if we’re keeping the margin cash in TLT? I guess maybe the call would make up for the losses in TLT and TLT would help in the next correction.
Would there be any benefit to keeping the margin in pure cash?
Method 3 starts looking good again as a hedge against the nest blowup.
A caveat would be an inflationary shock that sinks both stocks and bonds (and the 6% in long-VIX is not enough to compensate). It’s a risk and there is no way telling how this works out.
I do some covered calls with individual stocks in a retirement account. I don’t see why I would do more of that in my IB account. I already have equity beta through the puts, I have equity beta through the fixed income portfolio, so I don’t need more.
Hi Karsten,
I’m just checking that we’re talking about the method that utilises LEAPs?
I was just a bit confused because of the part where you mention VIX.
All good if you were. I will have to let your comment sink in a bit.
Sorry, I was confused. I misremembered #3 as the one with the VIX ETF. Yes, the LEAPS strategy would be timely to implement now that the VIX is low again!
Would keeping the rest of the cash in TLT still be a good idea in this environment? We can’t predict the future but it appears TLT would be a losing proposition in the rising rate environment.
Would you still say buying ATM LEAP would still be best? It offers the best downside protection.
However, if we were to sell shorter DTE calls to offset the LEAP, would it be better to buy a LEAP call that’s deep ITM (e.g. 70delta)? Would you sell a spread if you implemented this strategy?
I don’t have any specific simulations available but I really like the ATM LEAPS. Maybe Mr. Spintwig can run help us out?!:
Very insightful post!! I hope you will find the below post useful.
https://fundamentalmarket.wordpress.com/2020/08/17/weekly-market-newsletter-2/
What a difference a couple years makes. The leveraged stock + bond + vix strategy sure has been getting smoked this year!
Anything using leverage might have a tough time going forward as well with still relatively high valuations and borrowing rates at 4+% and likely going up with the Fed’s fight against inflation.
Yeah, that S/B correlation wasn’t a natural law after all. So, a lot of strategies in the risk parity space have gotten hosed.
Also shocking: The small allocation to VXX wouldn’t have helped much. The VIX didn’t even go very high this time around.
Maybe we should dust off the history books and see what worked well in the late 70s and early 80s un der Volcker, right?
Yeah, I saw that the 60/40 portfolio is having its worst year since 1936 if it doesn’t bounce back before the end of the year.
Even gold which did well in the 1970s hasn’t done well, just commodities and utilities.
Yup! Strange year this year. And it might get uglier next week when the FOMC meets.