How my collars are doing

[ChpBstrd]

Say the underlying you are writing options on pays a distribution.  Do you effectively give up the distribution as well as the upside?  Seems to me, if you were collaring something with a 5% yield, the cost of the collar would go up by the annualized amount of yield in your duration.  The price should rise as the distribution date draws near, meaning your cost to roll would go up if you were to roll near the x-div date.  If you wait until day after, your stock price theoretically falls by the amount of the distribution. Is this roughly a wash or is it an extra cost of collaring?

I am holding the stocks, so I get to keep the dividends. When I mention a X% maximum upside, that does not include dividends. The yield on QQQ is a mere 0.6% per year, which is meaningful but not a major consideration.

You are correct in your hunch that options prices must reflect dividends.

Ok, so if the hedge has a net debit, because of a high yield, but you hold through x-div, your net yield is theoretically ZERO. You have hedged the movement of the underlying price but surrendered the dividend to your counter parties!  Seems to me collaring BRK-B is much better than collaring SCHD or NOBL.  Maybe this is not important as if you are already taking profits off the table with distributions, and are investing in Aristocrats, you are already more conservative (and lower return) than going naked SPY.  So  you might already have all the 'hedge' you want (and have already paid for it! 

Correct me if I'm wrong, mixing the strategies is counterproductive though.

To clarify, if I hold 100 shares, -1 call, and 1 put, I still receive the dividend because I own the 100 shares. It's the price of the calls and puts we're talking about being affected by the arbitrage opportunity. Just wanted to clarify the language.
 
The price of the put tends to be higher before the ex-div date than after, because of demand for a nearly risk-free dividend or because put sellers fear the drop in the stock price after a large dividend. Kinda the same concept, from different perspectives, applicable to ITM or OTM options.

The price of the call tends to be lower before the ex-div date because there is a factor likely to cause the stock to fall: Namely the ex-div date. So when trading collars, you get less for selling your call. This is made up for by the receiving the dividend in the other hand.

Of course, stocks sometimes rise on the day after ex-div, so this is all reduced by probabilistic calculations. Yet these might be big issues for a person collaring on a month-to-month basis on a high-dividend stock like NLY or AGNC. They'd be selling the call lower and buying the put higher if they traded before the ex-div date. So they'd either have to constantly add cash into their options trades or accept lopsided pairs of strike prices. As you go further out in time, though, dividends make less and less of a difference because so many other factors could affect the stock price.

You might find an arbitrage opportunity, but it won't be one or two months out. Usually people think of a long stock plus synthetic short to capture risk-free dividends (basically a collar with the same strike prices that moves the exact opposite of the stock and has a delta near -1). I used to watch these closely, but have never had an order execute at advantageous pricing. This is also a situation where short call options are routinely exercised, even with lots of time remaining.

You might also find attractive, low-risk dividend plays. For example, you could write a synthetic short on NLY at the 20 strikes and 1/15/2027 expiration date for a debit of $-2.92 per share, or just buy the 20 strike put for $4.17. There are 21 $0.70 monthly dividends between now and then, so by pairing the synthetic short or protective put and the long shares, you could potentially obtain several very-low-risk dividends.

With the SS, I can tell you from experience that the short call options might be exercised early as they decay down to a certain critical level of time value, so you'd want to roll forward often. If a bot can turn the long call they bought from you into a profitable way to capture that dividend, then it will. The dividend-option plays I've run in the past have been gambles on getting executed or not. And with high-dividend stocks like NLY, you can endure the high risk of holding all month only to have your dividend stolen in the end when your short call is executed.

The protective put is more predictable, but you get something less than the stock's full yield because it is decaying and you get something less than a synthetic short's delta, so while it cannot be executed by a counterparty, it's also more risk in a different format. 

[ChpBstrd]

@MustacheAndaHalf interestingly ERN went through the effort to create a "better" "adjusted" CAPE in 2022 to account for recent decades changes in buybacks, accounting standards, taxes, and measurement issues.


The adjustments did lower CAPE significantly. The adjusted CAPE, when converted to CAEY, also had a tight relationship with the subsequent S&P500 ten year real returns in a more modern-era dataset from 2000-2012. This was an enormously volatile period of time, so one might expect more statistical noise. Instead ERN found:


In terms of too many datapoints... IDK. Using each month as a datapoint seems reasonable to me because prices can fluctuate a lot between months. We just need to be careful about extrapolating the number of data points into statistical significance or false precision, because we can manufacture as many data points as we want by just taking narrower slices of time. Such points would mostly fall in line with the trend visible here. I think the monthly data tell a compelling story of long-term returns being constrained by the valuation at time of purchase.

[Financial.Velociraptor]

@ChpBstrd

I have looked into the arbitrage trade on LEAPs for high yield stocks like MLP/BDC/mREIT.  The opportunity doesn't exist without a bot to trade it.  For you and I, the opportunity effectively doesn't exist.  That is, if you go with the EMH idea that today's price on a yield instrument is the best indicator of the price 1, 2, 5, 10, ?? years from now (without some additional shock), the cumulative dividends are reflective of the net debit to do an ATM synthetic short, less something very close to the discount that yields roughly what a treasury of the same duration is currently providing.  Without millions to spend to exploit pennies 20 times a day on hundreds of stocks, the retail investor is just as well off buying Treasuries at the duration that matches what period they want absolute safety for.  This leads me to think that collars are contra indicated for dividend stocks.  You still get most of the gain, without the downside but at the cost of both upside AND the yield. That is your yield after net debit is (EMH) close to zero and you are left without the security of the cash payment in the event the stock trades sideways.

[EscapeVelocity2020]

It would be really cool though, no matter how nerdy and inside baseball it might be, if Chp finds us an enduring edge on the Big Guys!  I'll jump through a few hoops waiting for low VIX or extending out a year on options, as long as my returns are guaranteed to enhance what I'm otherwise doing right now.  Heck, I might start a company and hand it over to my son, if this really pans out!

[ChpBstrd]

@ChpBstrd

I have looked into the arbitrage trade on LEAPs for high yield stocks like MLP/BDC/mREIT.  The opportunity doesn't exist without a bot to trade it.  For you and I, the opportunity effectively doesn't exist.  That is, if you go with the EMH idea that today's price on a yield instrument is the best indicator of the price 1, 2, 5, 10, ?? years from now (without some additional shock), the cumulative dividends are reflective of the net debit to do an ATM synthetic short, less something very close to the discount that yields roughly what a treasury of the same duration is currently providing.  Without millions to spend to exploit pennies 20 times a day on hundreds of stocks, the retail investor is just as well off buying Treasuries at the duration that matches what period they want absolute safety for.  This leads me to think that collars are contra indicated for dividend stocks.  You still get most of the gain, without the downside but at the cost of both upside AND the yield. That is your yield after net debit is (EMH) close to zero and you are left without the security of the cash payment in the event the stock trades sideways.

I too have searched for option/dividend El Dorado and not yet found it. For stocks like NLY (annual yield = 14.29%), you can synthetic short the shares at a much smaller debit than the upcoming quarterly dividend. The question is whether it makes sense for your counterparty to execute the call early. Any time an option is exercised early instead of just sold, the owner of the option sacrifices some time value. E.g. if the option has 50 cents of intrinsic value plus 25 cents of time value (total value: 75 cents), exercising it obtains only 50 cents but just selling it to someone else obtains 75 cents. Dividends are one of the only reasons an option holder might choose to early exercise, but once they've exercised, they hold a stock that is - on average - about to go down the same amount as the dividend. Right?

There's one theoretical issue I'm chewing on now. Markets should be efficient so that there is no risk-free large dividend on a protected put. However, markets also have to be efficient for collars or synthetic shorts, right? How can both be true at the same time? If it is a breakeven proposition to either buy a protective put or sell a covered put, then how do we account for the profit from selling a call on a stock that is probably going down due to the dividend?

Returning to NLY:
Share Price = 19.77
Jul18 20 Put = 1.32
Jul18 20 Call = 0.72
Dividend on 6/30 ex date = 0.70

A protective put would cost $21.09. On 7/1 your outcome would be an expected share price of 19.07, a dividend-in-hand of 0.70, plus the remaining value of the put with 17 days remaining on it. We can guestimate this value at around 0.48 because that is the mid price for the May 16 put, which has 15 days remaining and also has no dividend ahead. Putting these together, we obtain (19.07 + 0.70 + 0.48 =) $20.25. This is a net loss of $-0.84, so not recommended at these prices. However if you want to take the other side, short NLY and sell a put (a covered put), there appears to be an arbitrage opportunity over the next 2 months that would yield (0.84/21.09=) 4%. Maybe that makes sense depending on the short interest?

A shares + synthetic short would cost $20.37. On 7/1 you'd have: Expected share price of 19.07, dividend-in-hand of 0.70, the put worth an estimated 0.48, plus the short call worth an estimated -0.33. Total= $19.92, a loss of $-0.45. If you'd like to take the other side, shorting the stock and entering a synthetic long, you'd get a yield of (0.45/20.37=) 2.2%. This is probably very close to the short interest rate.

Perhaps a person/bot could earn the difference between these two strategies with a very complicated setup? IDK. Or perhaps the point is to earn the short interest instead of paying it? Or maybe you boost returns by letting that call go to expiration? And we haven't even touched on using different expiration dates to manage theta. It gets messy fast, which makes it appealing for simpletons like me to just do a covered call, decently OTM, and hope for the best. Even with an ultra-dividend stock like NLY, you can sell the 21 strike for about $0.28.

Also, I think the numbers for a collar / synthetic short look a LOT better if you go far out in time. A January 15, 2027 synthetic short pair of options at the 20 strike costs about $2.60, but the call has far more time value than the dividend in the next quarter, and will probably have more TV than the dividend in the quarter after that. So this seems unlikely to be assigned, compared to options a couple of months out.

In general, I don't like dividend stocks. Synthetic shorts can sell for a debit due to dividend arbitrage, as is the case for NLY, instead of a credit, as is the case for something like QQQ. If you're going to hedge your risk, why not aim for stocks that can grow? Dividend arbitrage is small potatoes in comparison to what QQQ has done. Plus, my QQQ/IWM short calls will never be assigned early due to dividends, so I don't have to watch it constantly.

I can use a collar to generate a "fake dividend" and pull ahead possible future gains* by just tilting my collar to generate a credit. I might do this if I was feeling bearish or was just shy of having enough money to afford a round lot of a stock. This technique might also be used to generate spending cash for bearish FIRE'ees during the critical early years.

*after all, that is what a dividend-paying company is doing by not reinvesting retained earnings.

[Financial.Velociraptor]

Perhaps a person/bot could earn the difference between these two strategies with a very complicated setup? IDK. Or perhaps the point is to earn the short interest instead of paying it? Or maybe you boost returns by letting that call go to expiration? And we haven't even touched on using different expiration dates to manage theta. It gets messy fast, which makes it appealing for simpletons like me to just do a covered call, decently OTM, and hope for the best. Even with an ultra-dividend stock like NLY, you can sell the 21 strike for about $0.28.

I think a "person" can not do this in 2025.  It was probably possible say in the 60s. Now there are ton of quant funds that have 100 million invested in hardware, software, and a super fast fiber connex racing each other to scalp what has now got be,  on average, pennies.  In an efficient market, there should always be an opportunity that is 'close' to the prevailing interest rate on US Treasuries of the same duration. 

I've done the out of the money call thing on NLY and AGNC etc.  The thing is the underlying is very sensitive to economic shocks.  Without the put, you are from time to time, take a huge haircut.  You are then writing your CC at a strike below your basis. Kind of like picking up pennies in front a steamroller.  Better to just suck it up and hold for the long slow crawl back up.  A better strategy is, I think to sell (far) out of the money puts monthly.  Let's call this a "lurking ambush predator" short put.  You are looking to earn a few pennies a month on the capital you tie up, hopefully keeping up with what you could get on monthly US gov debt. When the next liquidity crisis inevitably arrives, the price drops by half and you get assigned at say 60% of the previous spot.  You are now earning 20%+ while you wait plus covered calls that earn pennies at the strike of the underlying before the liquidity event.  You have a sweet distribution until the price recovers for something close to a double.  [Insert your best argument against 'market timing' here]

[MustacheAndaHalf]

@ChpBstrd - Middle graph first, of 2000-2012.  How many 10 year periods existed from 2000 to 2012?  Just three.  The 2001-2011 time frame had 90% overlap with 2000-2010 and 2002-2012.  Breaking 10 years of data into 120 months of data makes the overlaps worse: 10 year periods shifted 1 month have 99.2% overlap.  This graph does not show many separate data points, but data points that are nearly identical to each other.

@MustacheAndaHalf interestingly ERN went through the effort to create a "better" "adjusted" CAPE in 2022 to account for recent decades changes in buybacks, accounting standards, taxes, and measurement issues.
...

It is strange that graph has bigger gaps in the 1950s than in the 2000s.  The article does consider corporate tax rates and share buybacks, which is good.  In 2001, an accounting change to how assets were written off dramatically changed asset values:

"While FAS 142 may have introduced a more accurate accounting method, it also created an inconsistency in earnings measurements. Present values end up looking much more expensive relative to past values than they actually are. And the difference is quite dramatic. Adjusting for the accounting change would put the CAPE 10 about 4 points lower."
https://www.advisorperspectives.com/articles/2018/04/02/beware-of-the-misinterpretations-of-the-cape-ratio#:~:text=While%20FAS

Notice how in the 1900s, CAPE ratio averaged about 15, while in the past 25 years, it has averaged about 25.  If Larry Swedroe's estimate is accurate, 4 of those 10 points comes from accounting rule FAS 142, but there is still another 6 point gap left unexplained.  He explains another 0.5 to 1.0 gap with dividend payouts:

"To make comparisons between present and past values of the CAPE 10, any differences in payout ratios must be normalized. The adjustment between the 52% payout ratio (the average from 1954 through 1995) and the 35% payout ratio (the average from 1996 through 2017) corresponds to approximately a 1-point difference on the CAPE 10. Using the current payout ratio would lead to a smaller adjustment of about 0.5."


That article also points to another study, summarized below:

"Multiples such as D/P, P/E, and CAPE are useful when forecasting long-term returns, and largely useless when forecasting short-term returns. Given this mixed forecasting ability, the issue addressed in this article is whether these multiples can be used to devise successful asset allocation strategies in the sense of outperforming a simple static portfolio. The bulk of the evidence discussed here suggests that investors would be better off sticking with a simple 60-40 stock-bond portfolio."
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2594612

Poking around, I actually found the original study!
https://blog.iese.edu/jestrada/files/2015/10/MFAA.pdf

[ChpBstrd]

I think a "person" can not do this in 2025.  It was probably possible say in the 60s. Now there are ton of quant funds that have 100 million invested in hardware, software, and a super fast fiber connex racing each other to scalp what has now got be,  on average, pennies.  In an efficient market, there should always be an opportunity that is 'close' to the prevailing interest rate on US Treasuries of the same duration. 

I've done the out of the money call thing on NLY and AGNC etc.  The thing is the underlying is very sensitive to economic shocks.  Without the put, you are from time to time, take a huge haircut.  You are then writing your CC at a strike below your basis. Kind of like picking up pennies in front a steamroller.  Better to just suck it up and hold for the long slow crawl back up.  A better strategy is, I think to sell (far) out of the money puts monthly.  Let's call this a "lurking ambush predator" short put.  You are looking to earn a few pennies a month on the capital you tie up, hopefully keeping up with what you could get on monthly US gov debt. When the next liquidity crisis inevitably arrives, the price drops by half and you get assigned at say 60% of the previous spot.  You are now earning 20%+ while you wait plus covered calls that earn pennies at the strike of the underlying before the liquidity event.  You have a sweet distribution until the price recovers for something close to a double.  [Insert your best argument against 'market timing' here]

I actually like this play, although I'd keep my positions manageable. Definitely a steamroller, and it's debatable whether one would make future losses back more quickly through short puts or just by holding for dividends.

I've noticed in general, the market sort of "breathes" between a preference for certain, short-term money (value/dividends/low volatility) and uncertain, long-term money (growth/money-losing companies/high volatiltiy). During the recent correction, a lot of the high-yielding preferred stock I watch did not fall in value nearly as much as the tech companies. It is as if everyone's mood about the likelihood of payoffs in the distant future is constantly changing. Today, money is piling into the Nasdaq and coming out of gold and bonds, leaving dividend stocks in the dust. On another day, it'll be risk-off again, and the flow will reverse. At times when the Mr. Market is enthusiastic, it's time to sell him your growth stories and buy locked-in yields. The mood has to turn really pessimistic to get a deal on income producing equity though.

[ChpBstrd]

@ChpBstrd - Middle graph first, of 2000-2012.  How many 10 year periods existed from 2000 to 2012?  Just three.  The 2001-2011 time frame had 90% overlap with 2000-2010 and 2002-2012.  Breaking 10 years of data into 120 months of data makes the overlaps worse: 10 year periods shifted 1 month have 99.2% overlap.  This graph does not show many separate data points, but data points that are nearly identical to each other.

Fair point, but if we go back further in time, we start incorporating the regulatory/accounting/tax/payout/etc. features of the past. Then we have a weaker counterargument to the claim that "things are different now" and weaker support for ERN's claim that various edits to CAPE made it a better predictor of future 10 year returns. The monthly dots might be useful in the sense of offsetting seasonal effects that might occur if we, for instance, measured only each first day in January. "Can I retire this month?" was a fair question during the early 21st century.

Stocks were extremely volatile during this period, so my critique would be that the chart only shows the benefits of buying low, or retiring while stocks are low, during a period of seesaw action and high volatility. I believe the original scatterplot I posted from ERN, which can be found here, incorporates many more years of returns, but this is original CAPE, not "adjusted" CAPE.

ERN found that after the adjustments, adj-CAPE was significantly lower than CAPE and therefore the 4% rule looked a lot better in backtests using adj-CAPE in 2022:

(article says this chart is applying adj-CAPE)

Notice how in the 1900s, CAPE ratio averaged about 15, while in the past 25 years, it has averaged about 25.  If Larry Swedroe's estimate is accurate, 4 of those 10 points comes from accounting rule FAS 142, but there is still another 6 point gap left unexplained.  He explains another 0.5 to 1.0 gap with dividend payouts:

"To make comparisons between present and past values of the CAPE 10, any differences in payout ratios must be normalized. The adjustment between the 52% payout ratio (the average from 1954 through 1995) and the 35% payout ratio (the average from 1996 through 2017) corresponds to approximately a 1-point difference on the CAPE 10. Using the current payout ratio would lead to a smaller adjustment of about 0.5."

I suggest a lot of the remaining 5 to 5.5 gap reflects the increased liquidity and lower transaction costs investors have today, versus before the 1990s. We often forget how mutual funds often had four-figure "loads" and stocks could only be traded after paying three or four figure broker commissions to a person you talked to on the phone and whose office you visited to sign papers, only to wait a couple of days for the paperwork to go through. For most people, stocks were as illiquid as some highly restrictive hedge funds are today. It never made sense to trade less than a few thousand dollars at a time, because the loads/commissions/fees would eat up such a large percentage of the investment that it might take years to recover just the trading costs. Stocks had to have been discounted for these costs, which pushed most individual investors into bank CDs.

The high costs also led investors to prefer companies pay dividends for their income. It was simply not feasible to sell 5 shares of something to cover this month's bills. Dividends were a way of getting money out of companies without suffering trading fees. Thus companies paid higher dividends in the past, but this meant they could not reinvest as much of their retained earnings into the company or into new businesses. The modern growth company paying a zero dividend and posting negative earnings for its first decade was enabled by Etrade and the like in the 90s.

When stock ownership became democratized in the 1980s with 401k adoption and even moreso in the 1990s with online brokerages, stocks' values ceased to reflect the discount applied to the cost of getting in and out. They rose to reflect the removal of a huge cost.

ERN did not adjust for this variable. How would we do so, when the cost of trading as a percentage of the investment scaled up as the investment got smaller? Stocks are objectively more valuable in a world where you can hop in or out of ETFs containing hundreds or thousands of different companies for free at the click of a button, and where you can trade your $1,000 block as efficiently as if it was a $1M block.

[MustacheAndaHalf]

@ChpBstrd - Middle graph first, of 2000-2012.  How many 10 year periods existed from 2000 to 2012?  Just three.  The 2001-2011 time frame had 90% overlap with 2000-2010 and 2002-2012.  Breaking 10 years of data into 120 months of data makes the overlaps worse: 10 year periods shifted 1 month have 99.2% overlap.  This graph does not show many separate data points, but data points that are nearly identical to each other.

Fair point, but if we go back further in time, we start incorporating the regulatory/accounting/tax/payout/etc. features of the past. Then we have a weaker counterargument to the claim that "things are different now" and weaker support for ERN's claim that various edits to CAPE made it a better predictor of future 10 year returns. The monthly dots might be useful in the sense of offsetting seasonal effects that might occur if we, for instance, measured only each first day in January. "Can I retire this month?" was a fair question during the early 21st century.

Stocks were extremely volatile during this period, so my critique would be that the chart only shows the benefits of buying low, or retiring while stocks are low, during a period of seesaw action and high volatility. I believe the original scatterplot I posted from ERN, which can be found here, incorporates many more years of returns, but this is original CAPE, not "adjusted" CAPE.

ERN found that after the adjustments, adj-CAPE was significantly lower than CAPE and therefore the 4% rule looked a lot better in backtests using adj-CAPE in 2022:
...
(article says this chart is applying adj-CAPE)

(Note: I look at the graphs you provide carefully, but then exclude the image from my reply since it takes up so much space)

The key metric in the graph is CAPE above 20 or not.  But look at the earlier graph of CAPE and adjusted CAPE: over the past 25 years, CAPE has only been below 20 during the great financial crisis (2008-2009).  It might be helpful for ERN to pick a new dividing line to be more relevant to current CAPE ratios.

My original reply included a short paragraph acknowledging that two goals were in conflict: desire for many independent data points clashes with the relevance of earlier time periods using significantly different accounting for earnings.  Nothing to be done about that, since relative to 20 year measurements, it wasn't long ago.

Notice how in the 1900s, CAPE ratio averaged about 15, while in the past 25 years, it has averaged about 25.  If Larry Swedroe's estimate is accurate, 4 of those 10 points comes from accounting rule FAS 142, but there is still another 6 point gap left unexplained.  He explains another 0.5 to 1.0 gap with dividend payouts:

"To make comparisons between present and past values of the CAPE 10, any differences in payout ratios must be normalized. The adjustment between the 52% payout ratio (the average from 1954 through 1995) and the 35% payout ratio (the average from 1996 through 2017) corresponds to approximately a 1-point difference on the CAPE 10. Using the current payout ratio would lead to a smaller adjustment of about 0.5."

I suggest a lot of the remaining 5 to 5.5 gap reflects the increased liquidity and lower transaction costs investors have today, versus before the 1990s. We often forget how mutual funds often had four-figure "loads" and stocks could only be traded after paying three or four figure broker commissions to a person you talked to on the phone and whose office you visited to sign papers, only to wait a couple of days for the paperwork to go through. For most people, stocks were as illiquid as some highly restrictive hedge funds are today. It never made sense to trade less than a few thousand dollars at a time, because the loads/commissions/fees would eat up such a large percentage of the investment that it might take years to recover just the trading costs. Stocks had to have been discounted for these costs, which pushed most individual investors into bank CDs.

The high costs also led investors to prefer companies pay dividends for their income. It was simply not feasible to sell 5 shares of something to cover this month's bills. Dividends were a way of getting money out of companies without suffering trading fees. Thus companies paid higher dividends in the past, but this meant they could not reinvest as much of their retained earnings into the company or into new businesses. The modern growth company paying a zero dividend and posting negative earnings for its first decade was enabled by Etrade and the like in the 90s.

When stock ownership became democratized in the 1980s with 401k adoption and even moreso in the 1990s with online brokerages, stocks' values ceased to reflect the discount applied to the cost of getting in and out. They rose to reflect the removal of a huge cost.

ERN did not adjust for this variable. How would we do so, when the cost of trading as a percentage of the investment scaled up as the investment got smaller? Stocks are objectively more valuable in a world where you can hop in or out of ETFs containing hundreds or thousands of different companies for free at the click of a button, and where you can trade your $1,000 block as efficiently as if it was a $1M block.

I can't recall the last time I bolded anything from another person's post.  I emphatically agree lower commissions, lower sales loads, and decimalization all contributed.  I think figuring out how much requires checking with experts who know better.

Small cap stocks had wide big-ask spreads, which meant if you held them over short periods of time, you had a net loss from commissions.  The "small-cap factor" used in models of the market may be partially a liquidity premium - that people will pay more for large cap stocks they can sell at lower cost.

Before 2001, stock exchanges quoted prices in $0.0625 increments as a fraction: 1/16th.  After 2001, stock prices were always listed in $0.01 increments.  That paved the way for high-speed trading, which in turn provided greater liquidity overall (though an argument has been made during a crisis, it dries up).
https://www.investopedia.com/terms/d/decimalization.asp

I ultimately don't know how much lower costs pushed up prices and thus CAPE ratios, but it is definitely worth investigating.

[vand]

AQR have written a fairly devastating critique to this and similar option-based strategy which basically reinterates my doubts that they are better than just holding a well diversified portfolio.

https://www.aqr.com/Insights/Perspectives/Rebuffed-A-Closer-Look-at-Options-Based-Strategies

https://www.aqr.com/Insights/Perspectives/Buffer-Madness


quoting from the first article:

This note has focused on a single comparison: if you’re concerned with equity risk, are you better off a) using options or b) simply reducing your exposure to equities?  Obviously we believe, based on both theory and realized fact, that option b) is likely to be the better choice.
 
But investors have more choices than just these two. In fact, we think the best choice might be Option C: diversify better. Any strategy that offers positive risk-adjusted returns that are diversifying to equities is a candidate for improved portfolio outcomes.

[ChpBstrd]

AQR have written a fairly devastating critique to this and similar option-based strategy which basically reinterates my doubts that they are better than just holding a well diversified portfolio.

https://www.aqr.com/Insights/Perspectives/Rebuffed-A-Closer-Look-at-Options-Based-Strategies

https://www.aqr.com/Insights/Perspectives/Buffer-Madness


quoting from the first article:

This note has focused on a single comparison: if you’re concerned with equity risk, are you better off a) using options or b) simply reducing your exposure to equities?  Obviously we believe, based on both theory and realized fact, that option b) is likely to be the better choice.
 
But investors have more choices than just these two. In fact, we think the best choice might be Option C: diversify better. Any strategy that offers positive risk-adjusted returns that are diversifying to equities is a candidate for improved portfolio outcomes.

I enjoyed those articles. Thanks for finding them.

A couple of quibbles with their conclusions and the applicability:

1) I don't know the average expense ratio for the options-strategy funds the AQR writer used as comparisons, but between 0.5% and 1.5% seems like a good estimate. I don't pay most of these expenses, aside from typical commissions and fees of maybe $7 each time I reset a five or six figure position - maybe annually. Some funds will just buy another ETF, implicitly paying its expense ratio, and then charge their ER on top of it. That makes the compound ER is a little bit higher than what the fund is charging. I too have to pay ERs for the ETFs I hold, and this alone is enough to trail the index. But a fund of funds approach with layers of active management, regulatory, and marketing costs will definitely trail the index. Expense ratios are one reason I prefer to DIY it rather than buying one of the funds used in this comparison. I have yet to find an ETF that meets my needs as well as doing it myself. Even if I did find such a unicorn, it would be hard to justify the ER it would probably charge.

2) I'd be willing to bet many of the funds in the comparison, like the most-popular QYLD, employ a practice of writing at-the-money covered calls over short periods of time (i.e. one week or one month). This practice results in the call being ITM when the fund trades out of it about as often as it is OTM. They sometimes win cash and other times lose NAV. The only possible rationale here is to generate a large taxable dividend at the long-term expense of NAV. I say nope to that. If I want an annuity I'll buy an annuity.

3) Likewise, protective put funds typically "protect" for no more than a month at a time. That's a great strategy to pay the maximum on the time decay curve, while also providing no real protection against a realistic bear market that might involve a series of small losses each month for years. So, yes, I 100% agree with the authors' criticism here. I'm not satisfied with the strategy of any hedged ETFs out there today. Unless I'm missing something, no fund I've found is doing what I'd want them to do.

4) Collared ETFs tend to put the bumpers up at 5% to 10% away from the current price, and that's too close. BUFR, for example, is a fund of funds where each fund is aligned with one of 12 months, and the puts for each month are 10% OTM at the time the investment is started. The call's upside depends on market pricing, and is probably higher (For this service they charge a 0.95% ER and you get no dividend.). This fund is unappealing to me for writing/buying the options too close to the current price. I'm not hedging against a routine correction; I need to hedge against a SORR-level event. I'd prefer for both my put and my short call to expire worthless except in a minority of exceptional years, so I prefer a 20% max downside. In practice, I've found this reduces my volatility by almost half, which makes it a mystery to me how BUFR generates a beta of 0.64.

5) Options funds must exactly follow investing plans. They cannot decide to trade on a low-volatility or high-volatility day; it must occur per a calendar schedule regardless of whether it would be more prudent to do the opposite. They also cannot decide to drop hedges like I did last month with my IWM position. Buying puts during low volatility has a huge effect on the price you pay, and the financial outcome you receive. I at least exercise some control over what deals I will accept, rather than mechanistically trying to match some made-up index. I can see the results of doing so, because when I buy puts on low-vol days they have an upside bias and when I do the reverse I see a downside bias.

6) Basically, it's been an unusually good 5 years and an unusually good 10 years for stocks. There has been plenty of volatility, but massive economic stimulus, tax cuts, and persistent low unemployment have conspired to boost asset prices (and valuations) at the expense of a massive run-up in the U.S. national debt, and higher-than-expected inflation. The S&P500 had a total return of 113% over the past 5 years, which is far better than normal. I think the person who hedges today does not expect a repeat of that performance. If they did, they'd be going all-in on stocks or leveraged bets on the upside. While we're using past comparisons to predict future outcomes, why not prove that 100% Nvidia or Ethereum portfolios will be superior because they were superior in the past? A more interesting analysis might be to dot-plot the 5-year performance of all these funds against overlapping 5 year periods of market index performance over the past 50 years. I'm not saying the authors are cherry-picking data; they're just comparing a safety-first approach to being all-in during a massive bull run. Things could have turned out much differently.

There is an apples/oranges problem when comparing an option-hedged portfolio with a stocks+cash or stocks+bonds portfolio. The authors touch on this in their 2nd article covering criticisms. But the point is that a protective put position, for example, has a different functional line of possible returns than, say, a 60/40 portfolio, even if their betas are that same at this moment in time. The 60/40 portfolio can lose a lot more than the protective put position can lose, because its return function is a straight line, whereas the protective put hits a floor and can go no lower. As the authors note, the proper benchmark for a hedge fund is NOT a 100% stock portfolio, however they justify using stock+cash or stock+bond portfolios as the benchmark for option ETFs by referencing the funds' betas. Yet that beta is non-linear, and would change quickly if stocks took a hard fall. If, for example, stocks suddenly tumbled 20%, the beta for BUFR would drop because the puts would swing into the money. As I noted above, my positions using further-out strike prices seem to have less volatility than these funds, and I'm not sure why the funds are so volatile. Perhaps wider bid-ask spreads, low volumes, and investor confusion about the funds' exact holdings and NAV contributes? IDK. But what I do know is that a position with unlimited loss and upside is different than a position with limited loss and upside. From a SORR-avoidance perspective, the latter is more valuable because your retirement is less likely to collapse if invested that way.

Finally, how did they even have this conversation without talking about the Sharpe ratio? 
-----------------
So why isn't there an ETF that trades collars with one to two years till expiration, with flexibility to trade on low-volatility days, with flexibility to drop hedges if the market falls enough, with put strikes about 20% below the current price, and an ER below 0.3%? Probably such a thing would be impossible for either technical or marketing reasons.

Technical reasons might be violating the legal requirement of an ETF to track an index, no matter how contrived, or the risk of encountering low liquidity in the options market for far-OTM long-time-till-expiration options.

Marketing reasons might be the difficulty of explaining to people why they should be willing to accept up to 20% losses in a rare worst-case GFC-level scenario, in order to capture more of an even bigger upside much more frequently while enjoying much reduced volatility along the way and losing less, on average, per month, to time decay. That's no elevator pitch, especially considering how I've never met a person in my daily life who knows what SORR is. Instead, everyone is obsessed with one-to-five year outperformance of some obscure index, if not simply chasing performance.

[MustacheAndaHalf]

@ChpBstrd - "Equity Buffer" ETFs provide a buffer from downside moves, while capping upside.  But it requires buying in the start month, and holding 12 months until the next reset period.  Between those dates, the market decides how to pro-rate the protection and cap.  The upcoming $BJUN ETF limits upside to +16.64% and absorbs ("buffers") the first 9% of downside.  If the market is down 20%, this ETF should be down 11%.  Expense ratio of 0.79%.

https://www.innovatoretfs.com/etf/default.aspx?ticker=bjun

I considered them a year or so ago, but decided the upside cap wasn't worth it for me.

[ChpBstrd]

IWM:
I exited this collar on 4/10/25 for a profit on the options of $7.48 per share, because IWM had fallen more than 20% from its peak. As the historical statistics predicted, the shares rose after that correction. IWM is up 8.65% since I dropped my hedges, so I enjoyed the full effect of that upswing. I expect to re-enter a long term collar after IWM has recovered to its ATH, around +10.6% from here.

I sold a couple of 1-3 day covered calls on my IWM out of boredom since then, but didn't make any significant money. I've decided to cease writing CC's based in part on experience and in part on insights from this article.

I do not see the article's results as damming collars too, because at the >1 year durations I'm using, it is a lot easier to guess the range where the market will land unless there is a bear market, because the downsides of covered calls are covered by adding a put and making it a collar, and because the theoretical basis of what I'm doing is different, in terms of the authors' discussion of earning the volatility risk premium (a costless collar both earns the VRP by selling the call, and spends the VRP by buying the put). However as we'll see with my QQQ positions, I'm living with some regret over my CC's.

QQQ:
I had my finger on the trigger, but did not exit my QQQ collars in April even as losses danced around my 20% threshold. This is regrettable in hindsight because QQQ rose sharply from that point. At the time I didn't have any way of knowing additional losses were not on the way, so I stuck with a strict interpretation of my investment policy statement.

However, I then did something stupid with two of my QQQ positions. I ignored my IPS, got greedy, and tried rolling down some of my calls to a lower strike price. Doh! Should have left my collars intact. That's the whole point of making it a long-term strategy!

So my simple collar strategy branched out into three different positions:

• Still Collared: 400 shares in an IRA at 409.78 and 569.78 expiring 6/20/25. I originally entered this collar 10 months ago. The put is 21.2% out of the money. The call is 9.6% OTM. I don't see either outcome happening in the next 22 days, so I'll just hold these options positions till expiration rather than throwing any money at an early exit (it would cost a whole six dollars to exit, lol, but that's a free taco just for waiting). This does leave me exposed though, because the combined delta is only -0.025. Basically there's not much difference between being in this collar and not being in this collar. I'm fine with that, because I see our current position as being at the beginning of a recovery from the March-April correction that could last many more months. I try not to think about the thousands of dollars these options might have netted had I exited them at the same time I did IWM. Yet I didn't lose anything and enjoyed lower volatility along the way, so it worked as planned and the most likely / expected outcome was achieved.
• Still Collared:200 shares in a taxable account at 434.78 and 574.78, expiring 12/19/2025. The put is 16.4% OTM and the call is 10.6% OTM. The combined delta is 0.46. I will probably roll this in June or July, but for now I'm happy with the delta.
• Covered Call Limbo at 520: 300 shares in a Roth IRA where I sold the 409.78 6/20/2025 put option for $100 on 5/20, converting the position to a covered call. I had earlier done something dumb with the short call. On 3/31 I rolled it down from the 569.78 strike to the 500 strike, for a lousy $2,035 credit. I was feeling pessimistic and underestimated how quickly the Chinese would break Trump in negotiations. Then on 5/16 I tried rolling up and out, trading my 500 strike for a 520 strike and moving maturity out from 6/20 to 9/19. That transaction generated a net credit of about $66. So now I have an at-the-money covered call expiring almost 4 months out. Yet this is not dead money. First, the short call has about $28.36 of time value that will erode to nothing by mid-September. That's 5.77% of the position that I'll receive as a gain if I just hold and do nothing, because that's the time value left on the call. 5.77% is 4 months isn't a bad return. Plus I'm well-hedged with a delta of 0.556. The downside, though, is the much higher risk of having to buy back in at a higher price. I should have just let my original collar work out like the position above.
• Covered Call Limbo at 500: I repeated the above mistake for 400 shares in another IRA. Except I've not yet done anything about the 500 call, which expires 6/20/2025. The delta is up to 0.766 and there is currently $5.00 per share of time value to erode over the next 3 weeks, offering a 1% return in that time. The short call is decaying at a rate of 23 cents per day, per share. So the whole position is harvesting $92 per business day in time decay. Because of the high delta, this position isn't moving much but it's also my most solid hedge against another dip. In fact, this very solid hedge makes me more comfortable with the risk of my very-lightly hedged positions. It's possible for these calls to expire OTM if QQQ falls 4% over the next 3 weeks. Still, it burns to see a $-7,350 loss on this particular option. So again, I punished myself for deviating from the plan.

It's fair to ask if I should regret converting some of my QQQ collars to covered calls. I'm earning double-digit annualized returns on time decay, enjoying some downside protection, benefiting from a decline in implied volatility, and to some extent at the same delta where I was hedging with collars. My regret comes down to an intuition that we're probably beginning a long recovery from the March-April correction and a worry that I'll end up breaking apart my nice 100-share lots of QQQ and ending up with un-hedgable odd lots.

Yet it costs money to exit a covered call, and that money could only come out of other investments, like my recently accumulated gold, UK stock, and Brazil stock ETF positions or an 8% yielding preferred stock. Would the opportunity cost of selling these assets be less than the opportunity cost of QQQ rising from here, multiplied by some fraction?

[vand]

AQR have written a fairly devastating critique to this and similar option-based strategy which basically reinterates my doubts that they are better than just holding a well diversified portfolio.

https://www.aqr.com/Insights/Perspectives/Rebuffed-A-Closer-Look-at-Options-Based-Strategies

https://www.aqr.com/Insights/Perspectives/Buffer-Madness


quoting from the first article:

This note has focused on a single comparison: if you’re concerned with equity risk, are you better off a) using options or b) simply reducing your exposure to equities?  Obviously we believe, based on both theory and realized fact, that option b) is likely to be the better choice.
 
But investors have more choices than just these two. In fact, we think the best choice might be Option C: diversify better. Any strategy that offers positive risk-adjusted returns that are diversifying to equities is a candidate for improved portfolio outcomes.

I enjoyed those articles. Thanks for finding them.

A couple of quibbles with their conclusions and the applicability:

1) I don't know the average expense ratio for the options-strategy funds the AQR writer used as comparisons, but between 0.5% and 1.5% seems like a good estimate. I don't pay most of these expenses, aside from typical commissions and fees of maybe $7 each time I reset a five or six figure position - maybe annually. Some funds will just buy another ETF, implicitly paying its expense ratio, and then charge their ER on top of it. That makes the compound ER is a little bit higher than what the fund is charging. I too have to pay ERs for the ETFs I hold, and this alone is enough to trail the index. But a fund of funds approach with layers of active management, regulatory, and marketing costs will definitely trail the index. Expense ratios are one reason I prefer to DIY it rather than buying one of the funds used in this comparison. I have yet to find an ETF that meets my needs as well as doing it myself. Even if I did find such a unicorn, it would be hard to justify the ER it would probably charge.

2) I'd be willing to bet many of the funds in the comparison, like the most-popular QYLD, employ a practice of writing at-the-money covered calls over short periods of time (i.e. one week or one month). This practice results in the call being ITM when the fund trades out of it about as often as it is OTM. They sometimes win cash and other times lose NAV. The only possible rationale here is to generate a large taxable dividend at the long-term expense of NAV. I say nope to that. If I want an annuity I'll buy an annuity.

3) Likewise, protective put funds typically "protect" for no more than a month at a time. That's a great strategy to pay the maximum on the time decay curve, while also providing no real protection against a realistic bear market that might involve a series of small losses each month for years. So, yes, I 100% agree with the authors' criticism here. I'm not satisfied with the strategy of any hedged ETFs out there today. Unless I'm missing something, no fund I've found is doing what I'd want them to do.

4) Collared ETFs tend to put the bumpers up at 5% to 10% away from the current price, and that's too close. BUFR, for example, is a fund of funds where each fund is aligned with one of 12 months, and the puts for each month are 10% OTM at the time the investment is started. The call's upside depends on market pricing, and is probably higher (For this service they charge a 0.95% ER and you get no dividend.). This fund is unappealing to me for writing/buying the options too close to the current price. I'm not hedging against a routine correction; I need to hedge against a SORR-level event. I'd prefer for both my put and my short call to expire worthless except in a minority of exceptional years, so I prefer a 20% max downside. In practice, I've found this reduces my volatility by almost half, which makes it a mystery to me how BUFR generates a beta of 0.64.

5) Options funds must exactly follow investing plans. They cannot decide to trade on a low-volatility or high-volatility day; it must occur per a calendar schedule regardless of whether it would be more prudent to do the opposite. They also cannot decide to drop hedges like I did last month with my IWM position. Buying puts during low volatility has a huge effect on the price you pay, and the financial outcome you receive. I at least exercise some control over what deals I will accept, rather than mechanistically trying to match some made-up index. I can see the results of doing so, because when I buy puts on low-vol days they have an upside bias and when I do the reverse I see a downside bias.

6) Basically, it's been an unusually good 5 years and an unusually good 10 years for stocks. There has been plenty of volatility, but massive economic stimulus, tax cuts, and persistent low unemployment have conspired to boost asset prices (and valuations) at the expense of a massive run-up in the U.S. national debt, and higher-than-expected inflation. The S&P500 had a total return of 113% over the past 5 years, which is far better than normal. I think the person who hedges today does not expect a repeat of that performance. If they did, they'd be going all-in on stocks or leveraged bets on the upside. While we're using past comparisons to predict future outcomes, why not prove that 100% Nvidia or Ethereum portfolios will be superior because they were superior in the past? A more interesting analysis might be to dot-plot the 5-year performance of all these funds against overlapping 5 year periods of market index performance over the past 50 years. I'm not saying the authors are cherry-picking data; they're just comparing a safety-first approach to being all-in during a massive bull run. Things could have turned out much differently.

There is an apples/oranges problem when comparing an option-hedged portfolio with a stocks+cash or stocks+bonds portfolio. The authors touch on this in their 2nd article covering criticisms. But the point is that a protective put position, for example, has a different functional line of possible returns than, say, a 60/40 portfolio, even if their betas are that same at this moment in time. The 60/40 portfolio can lose a lot more than the protective put position can lose, because its return function is a straight line, whereas the protective put hits a floor and can go no lower. As the authors note, the proper benchmark for a hedge fund is NOT a 100% stock portfolio, however they justify using stock+cash or stock+bond portfolios as the benchmark for option ETFs by referencing the funds' betas. Yet that beta is non-linear, and would change quickly if stocks took a hard fall. If, for example, stocks suddenly tumbled 20%, the beta for BUFR would drop because the puts would swing into the money. As I noted above, my positions using further-out strike prices seem to have less volatility than these funds, and I'm not sure why the funds are so volatile. Perhaps wider bid-ask spreads, low volumes, and investor confusion about the funds' exact holdings and NAV contributes? IDK. But what I do know is that a position with unlimited loss and upside is different than a position with limited loss and upside. From a SORR-avoidance perspective, the latter is more valuable because your retirement is less likely to collapse if invested that way.

Finally, how did they even have this conversation without talking about the Sharpe ratio? 
-----------------
So why isn't there an ETF that trades collars with one to two years till expiration, with flexibility to trade on low-volatility days, with flexibility to drop hedges if the market falls enough, with put strikes about 20% below the current price, and an ER below 0.3%? Probably such a thing would be impossible for either technical or marketing reasons.

Technical reasons might be violating the legal requirement of an ETF to track an index, no matter how contrived, or the risk of encountering low liquidity in the options market for far-OTM long-time-till-expiration options.

Marketing reasons might be the difficulty of explaining to people why they should be willing to accept up to 20% losses in a rare worst-case GFC-level scenario, in order to capture more of an even bigger upside much more frequently while enjoying much reduced volatility along the way and losing less, on average, per month, to time decay. That's no elevator pitch, especially considering how I've never met a person in my daily life who knows what SORR is. Instead, everyone is obsessed with one-to-five year outperformance of some obscure index, if not simply chasing performance.

thanks ChpBstrd - I really do appreciate the detailed rebuttal.. even if I'm still far from convinced.

Without wanting to highlight specific points, they critical the flaw which AQR highlighed in all these strategies is pointed out in the original article, and it is about the misunderstood protective power of puts, for which the profit-window is narrower than is widely assumed.  Puts are designed to profit if the market moves to a certain level by a certain date - the path it takes matters, which is why you can lose protection with eg 10% downside puts even with the market down eg 16% if it only goes down 8% one month, then 8% the next, so the put which are rolled over monthly never reach in the money.  Obvious this is from a monthly rollover, but the same principle applies to any timeframe you are considering.

You say that you are trying to protect against >20% SORR level of downside rather than just regular volatility.. but a large crash isn't a necessary condition for SORR - it can unfold more gradually with stubborn sideways market too.

[ChpBstrd]

You say that you are trying to protect against >20% SORR level of downside rather than just regular volatility.. but a large crash isn't a necessary condition for SORR - it can unfold more gradually with stubborn sideways market too.

This is a great point. Historical pricing charts show relatively flat markets for long periods of time in the 1970s, 1900s, and 1890s, and a long highly-volatile back-to-where-we-started market between 1999 and about 2011.

Arguably most returns would have come from dividends during such periods, and covered calls might have been a winning strategy if they had been possible. Returns would have usually landed between the upper and lower range of the sort of collars I've been using, so they'd have no effect if they were free.

If you're making 4% or 5% or higher withdraws in a market that is going nowhere for several years at a time, the only help a long-duration, wide-strikes collar is going to provide is the courage to stay in stocks. It is a defensive strategy that does not typically score points on its own, although I like my rule of exiting the options when stocks are down at least 20%.

...the flaw which AQR highlighed in all these strategies is pointed out in the original article, and it is about the misunderstood protective power of puts, for which the profit-window is narrower than is widely assumed. 

I can also read this as a critique of the protective put strategy. I.e. if you pay $5 for a put at the 100 strike, your potential loss at expiration is not all the way down to $100; it's all the way down to $95. This is because you can lose stock value all the way down to the strike price PLUS whatever you paid for the put. It's lazy thinking in this example, to say "I'm protected down to $100" and I've been guilty of that fallacy in the past.

So I like collars sold at near breakeven because they do not involve a net payment for time value. Your maximum up and maximum down at expiration are the strike prices, which simplifies calculations and thinking. I think the quote above was talking about time, but read a different way it can talk about time decay, which occurs more quickly as a percentage of the option premium the closer the option gets to expiration.

Puts are designed to profit if the market moves to a certain level by a certain date - the path it takes matters, which is why you can lose protection with eg 10% downside puts even with the market down eg 16% if it only goes down 8% one month, then 8% the next, so the put which are rolled over monthly never reach in the money.  Obvious this is from a monthly rollover, but the same principle applies to any timeframe you are considering.

In my mind, real SORR protection comes from mitigating the effects of realistic disaster scenarios where the 4% rule failed (or will probably fail) in the past. These situations almost always involve back-to-back double-digit declines in stock price returns and can be thought of as sequences. Example S&P500 price return sequences include:

Dot-Com Bubble Sequence:
2002    -23.37%
2001    -13.04%
2000    -10.14%

70's Oil Embargo and Inflation Sequence:
1974    -29.72%
1973    -17.37%

Early Great Depression Sequence:
1932    -14.78%
1931    -47.07%
1930    -28.48%
1929    -11.91%

So as we try to imagine what it would take for a stock-heavy portfolio to survive such sequences, questions come to mind such as "how can I put a floor on losses?" or "how can I profit from a dramatic increase in volatility?" or "how could I set up a hedging algorithm that absorbs hits and then gets out of the way to enable gains?" In my mind, this could only be done on >1 year timeframes. A more ideal hedging strategy would mitigate damage across 3-4 year timeframes, but LEAPS options only go out a little over 2.5 years.

Still, we can wonder what the SWR for these periods would have been had the losses for the first two years been limited to, say, -25% total? Or if the maximum loss any given year was -15% or -20%. Given that the 4% rule almost survived these sequences, it seems like such an adjustment might have made the difference.

Then there are the one-off bad years, like 2008 (-38.49%) or 1937 (-38.59%) which were followed by strong gains, but still represent SORR for retirees forced to sell their assets. Mitigating the damage from just one such event could make the difference.

I mean, maybe all the one-month collar or put strategies are dealing with delta the same as I am, and maybe because they're so close to the money the ratcheting down process isn't as big a deal as it seems. But they're forced to pay for extreme volatility when they roll in the midst of crisis and they're set up to miss out on fast recoveries like we saw in April and May of this year. If they can't perform well enough in a normal market, what kind of protection should we expect in an actual SORR event?