November 10, 2021
Welcome to a new post in the Put Option Writing Series. My blogging buddy Spintwig volunteered to perform another backtest simulation. If you remember from Part 5, he simulated selling 5-delta and 10-delta put options going back to 2018. He now added 18 more months of returns to go back to September 2016. In the end, I will also compare my live results with the simulated returns and point out why my live trading achieved even slightly better results.
Mr. Spintwig, please take over…
* * *
Thank you BigERN (can I call you Dr. K?) for another opportunity to collaborate and add to the body of research that supports what is colloquially known as the “BigERN strategy.”
Part 8 of the options trading series is a 2021 update that discusses, among other things, premium capture, annualized return and the idea of lowering leverage while increasing delta.
Let’s throw some data at the idea of trading a higher delta at a lower leverage target and see how metrics like premium capture, CAGR, and max drawdown are impacted. As an added bonus, I’ve obtained SPX data that can facilitate a Sept 2016 start date for this strategy. This gives us an additional 18 months of history vs the SPY data that was used in Part 5.
For the benchmark, we’ll use total return (i.e. dividends reinvested) buy/hold SPY (S&P 500) and IEF (10Y US Treasuries), rebalanced annually, in the following configurations:
- 100 SPY / 0 IEF
- 80 SPY / 20 IEF
- 60 SPY / 40 IEF
Let’s dive in…
Strategy: Short Put
Days Till Expiration: 2 DTE (Mon, Wed); 3 DTE (Fri)
Start Date: 2016-09-01. End Date: 2021-10-29
Positions opened per trade: 1
Entry Days: Mon, Wed, Fri, at 3:46 pm Eastern Time (i.e., 14 minutes before market close)
- 5 delta +/- 4.5, closest to 5
- 10 delta +/- 5, closest to 10
- 5D short put
- 10D short put
Trade Exit: Hold till expiration
Max Margin Utilization Target: 20%-100% (1x to 5x leverage) in 20% increments
Max Drawdown Target: 99% | account value shall not go negative
- Margin requirements are always satisfied
- Margin calls never occur
- Margin requirement for short put positions is 20% of notional
- Early assignment never occurs [ERN: it won’t, as these are European Options]
- Prices are in USD
- Prices are nominal (not adjusted for inflation)
- All statistics are pre-tax, where applicable
- Margin collateral is invested in 3mo US treasuries and earns interest daily
- Assignment P/L is calculated by closing the ITM position at 3:46pm ET the day of expiration / position exit
- Commission to open, close early, or expire ITM is 1.32 USD per contract
- Commission to expire worthless is 0.00 USD per contract
- Slippage is calculated according to the slippage table
- Starting capital for short option backtests is adjusted in $1000 increments such that max margin utilization is between 80-100%, closest to 100%, of max margin utilization target
- Starting capital for long option backtests is adjusted in $1000 increments such that max drawdown is between 80-100%, closest to 100%, of max drawdown target
- For comprehensive details, visit the methodology page
Let’s look at the results
Starting capital between leverage targets is non-linear (eg: 2x leverage isn’t exactly half as much as 1x) due to the varying amounts of interest earned on margin collateral and other strategy mechanics.
Increasing the delta target from 5 to 10 enables the strategy to execute within the backtest parameters with a smaller amount of starting capital.
Average margin utilization is roughly the same between both delta targets.
Hindsight bias was used to maximize Reg-T margin utilization for each strategy. This allows a “best case” scenario for the option strategy to outperform the benchmark. Also displayed is the date on which each strategy experienced maximum margin utilization.
Premium capture for 10-delta strategies is materially lower than for 5-delta strategies. Nevertheless, a 37% profit margin on premium sold isn’t a bad deal.
The 10-delta strategy experienced a lower win rate vs the 5-delta strategy. [ERN: Not a huge surprise here because the strikes are higher and thus the 10-Delta options are more likely to end up in the money.]
Despite the smaller amount of premium capture per above, the 10-delta strategy outperforms the 5-delta strategy with regard to average monthly P/L.
Also displayed is the best and worst monthly return for each strategy:
The max drawdown for the 10-delta strategies is higher than the 5-delta strategies.
Maximum Drawdown Duration
10-delta strategies experienced greater max drawdown durations than 5-delta strategies.
Compound Annual Growth Rate
The 10-delta strategies outperformed the 5-delta strategies with regard to CAGR.
10-delta strategies experienced greater volatility than 5-delta strategies.
5-delta strategies outperformed 10-delta strategies with regard to risk-adjusted return. The greater the leverage the lower the risk-adjusted returns.
Percent of Profit Spent on Commissions
The 10-delta strategies generated and retained more premium than the 5-delta strategies. Thus, commissions are a smaller proportion of the total return. [ERN: for clarification, the numbers displayed are the percent of the gross option profits lost to commissions, not the percentage of the account value!]
10-delta strategies outperformed 5-delta strategies with regard to total return. [ERN: again, as expected, but note that the higher return also comes at the cost of much higher volatility!]
Regarding premium capture, there is a material reduction when going from 5-Delta to 10-Delta. Nevertheless, the total return increases. A smaller amount of premium is retained from a much larger “pie.”
When we look at risk-adjusted return, the lower delta strategy has a materially higher Sharpe Ratio. Reviewing the max drawdown and max drawdown duration between the two option strategies supports this observation.
However, the best and worst monthly returns are about the same. For example, the worst monthly return of a 3x leveraged 5-delta strat vs a 2x leveraged 10-delta strategy is similar.
The worst monthly return is the best proxy available for measuring the impact of left-tail events. Understandably, it leaves much to be desired and doesn’t truly capture the impacts of rapid shocks like flash crashes or circuit-breaker-tripping moves.
Dialing up the delta while reducing the leverage has yielded a bumpier ride. Structurally, it guarantees a less severe left-tail event since there is less leverage in play.
Back over to you Dr. K —
When I worked at BNY Mellon, some people called me Dr. J, like the basketball player. But Dr. K works just as well! But in any case, thanks, Mr. Spintwig for volunteering to run this new extended case study and simulation. Nice to see that the February 2018 blowup didn’t pose any problems to your simulations. Quite the opposite, by including the late 2016 and all of 2017 we even increased the overall Sharpe Ratio. For example, the 5-Delta, 3x leverage strategy Sharpe Ratio was 1.71 in the 2018-2020 simulation, but 2.12 in the 2016-2021 simulation.
How do the simulations compare with my live returns? Well, here’s something we don’t see too often: my live returns look substantially better than the simulated returns. I achieved a Sharpe Ratio of just about 4.0, which is significantly higher than in the simulations. It’s not because I had much higher returns, but much lower risk. My average strategy return (11.9%) lines up roughly with the Spintwig 5-Delta, 4x leverage return (11.1%), or the 10-Delta 3x leverage (10.8%).
Then, what caused the lower risk and higher Sharpe Ratio? There are (at least) three reasons:
1: Market Timing, a.k.a. dumb luck!
If we look not just at the overall returns but also the returns during the subperiods, notice that during the relatively calm period up to January 2018, I vastly outperformed the simulated strategies with 5-Delta puts and even the 10-Delta put with 3x leverage. Afterward, my average returns were much more modest. And that’s easy to explain. The early period still covers my accumulation phase and I took much larger risks. Then in February 2018, I announced my plans to retire and we also received the proceeds of our San Francisco condo sale and a large portion of that went straight into the IB account. I took the put writing risk down a notch in light of the much larger account size and the realization that I’d have to live off the IB account proceeds. That was brilliant market timing. Or, well, mostly dumb luck, because I ran the strategy full-throttle during the very profitable time in 2016 and 2017, but then treaded a lot more cautiously right when volatility hit.
During the last year and even 3 years, I mostly underperformed the aggressive fixed-delta, fixed leverage strategies that Spintwig simulated here today. And I’m totally happy with that because even at that modest risk target, I generate enough retirement income to fund our lifestyle.
2: Targeting a fixed gross put writing revenue as opposed to a fixed Delta lowers the monthly return volatility
Take a look at the chart where I plot my monthly returns against the monthly returns of the 5-Delta, 4x leverage returns:
Notice the differences?
- As we already know, I outperformed the Spintwig strategy in 2016/17 because I ran it with a higher risk tolerace.
- During normal times, i.e., months when you simply make your full gross option premium, my returns are extremely range-bound, with very little volatility. Look at the almost straight line after the pandemic. And that’s by construction becuase I target a fixed income per month. In contrast, Spintwig’s simulation had significantly more volatility since then, even some losses. Actually, post-pandemic I was running my strategy with a very low Delta, all the way down to 2. Though, my Delta has since moved up again to about 4 because I also sell fewer contracts, as I outlined in Part 8 of the series. With less leverage you have to increase the Delta to keep the income the same!
- Right after the February 2018 volatility spike, the fixed-Delta strategy got hammered again in March. The worst monthly return in the simulation! In contrast, my losses were limited because I targeted my income stream, which called for a lower Delta and less leverage. I still lost another 1% in March 2018, but the loss wasn’t as painful as in the simulation with a 4x leverage and a fixed 5-Delta.
3: Dumb luck (again!) in Q1 of 2020!
The one time when I deviated from the “fixed gross put-writing revenue” target was the pandemic bear market. After suffering two smaller losses in January 2020, a very stinging loss on February 24, and a close call later that week I saw a significant volatility spike. Sure, I could have continued to write options with a $1.00 premium Monday to Wednesday. But the implied volatility was so high and the option premiums so rich, I sold options with a 1-Delta or 2-Delta that still yielded more income than I budgeted. And I also did a few additional intra-day trades and supplemental trades on Tuesdays and Thursdays when the current options had already lost most of their value. See Part 4 of the series.
So, in summary, the improvement in the Sharpe Ratio came mostly through the lower volatility and a little bit of luck about timing my target risk. I don’t want to make a big fuss out of that. No need to spike the football. I would have been happy with the Spintwig simulation results as well.
Monthly vs. Daily returns
With a Sharpe Ratio of 4, i.e., roughly 2.5% annualized risk, and 10% annualized excess returns, one might be tempted to lever up this strategy. Why not target 10% risk and 40% annual (excess) returns? Retire after maybe 2-3 years of accumulation and then run this strategy with a 30% safe withdrawal rate?! All of our worries are solved, right?
Well, not so fast! Hidden in the low monthly standard deviation is the fact that there is significant intra-month volatility. Plotting the monthly and daily cumulative returns, we can see that during some of the stress periods (Q1 2018, Q4 2018, and Q1 2020) you had some unpalatable intra-month vol that is not visible in the monthly return numbers. If you had levered up the strategy much more you might have faced margin calls, forced liquidations, and you might have locked in catastrophic losses. In contrast, in my live strategy with relatively modest leverage, I recovered most losses rather swiftly.
So, again, a warning to all current and prospective option strategy fans: Don’t fall for rosy monthly Sharpe Ratios and don’t use too much leverage. I have written about two examples where option writing went very wrong because people were sloppy with their risk management:
To hedge against a potential “Black Swan” event, you want to have a sizable safety cushion.
So, I’m totally fine with a 10% annualized return. And I should also stress that I don’t compare my 10% put writing return to the 17% equity return over the last 5 years. I need to compare my strategy returns to a 1.5% to 2% bond yield. As I outlined in Part 8 of the series (see section 2 “How does the Put Selling Strategy fit into our overall portfolio?”), I have plenty of equity and real estate exposure in the other accounts already. Instead of keeping 25-30% of safe assets like bonds, money markets, TIPS, I Bonds, etc. with lousy return prospects, I prefer to take a little bit of extra risk. Writing puts suddenly looks very attractive again! Also, keep in mind that the 10% return is in addition to the interest and dividends from the fixed income portfolio.
So, in any case, I just wanted to point out the slight discrepancy between the simulations and the live returns. I don’t want to make too much out of the difference. Mostly because it’s due to lower realized risk in my live returns, but that’s not something I want to bank on because there’s still a lot of hidden black swan risk that – so far at least – has never materialized. But that risk is always on my mind!
Thanks for stopping by today. Please make sure you check out the other parts of this series:
- Trading derivatives on the path to Financial Independence and Early Retirement
- Part 1 – Intro
- Part 2 – Extended Intro
- Part “2.5” – Trading like an Escape Artist: October 2018 update
- Part 3 – Strategy details as of 2019
- Part 4 – Surviving the 2020 Bear Market!
- Part 5 – A 2018-2020 backtest: Guest Post by “Spintwig” (plus a quick update on last week’s volatility)
- Part 6 – A 2018-2021 backtest with different contract sizes: Guest Post by “Spintwig”
- Part 7 – Careful when shorting long-dated options!
- Part 8 – A 2021 Update
- Part 9 – A 2016-2021 backtest: Guest Post by “Spintwig”
Title picture: pixabay.com (or maybe it’s Big ERN’s secret trading floor???)