Corner of Berkshire & Fairfax Message Board
General Category => Strategies => Topic started by: Ross812 on July 30, 2018, 11:15:28 AM

My father approached me to look into into market indexed annuities. These annuities essentially give the holder a percentage of market upside up to a cutoff and protect the holder from any market downside. In this case, it was 70% of market upside up to 12% with no downside. I looked into the annuities and found out they were essentially a repackaged market linked CDs (MLCDs) sold by banks in 250k to 1M dollar increments. The MLCDs had better terms, so i figure the insurance company is keeping a bit of the upside for their trouble plus adding some heavy origination fees for the annuity to pay their salesmen.
I started looking if there was a way to accomplish the same thing on more favorable terms myself. The objective was to provide capital preservation while still participating in the market. I came up with a strategy of laddering 1 to 4 yr CDs, AA, and AAA debt offered on the secondary market, then using the interest proceeds to buy long dated out of the money LEAP options on the the S&P 500 index through XSP. I chose XSP over SPY because the options are European Style (noncallable), settled to cash, and the underlying index is adjusted for dividends.
To mirror the MLCD my father was looking into you would essentially buy ATM calls and sell OTM calls 12% above the market price.
I created a spreadsheet to backtest the strategy and found it is far better to by options further OTM to lever the 500k. Here is the distribution of SP500 returns since 1825:
(http://www.mymoneyblog.com/wordpress/wpcontent/uploads/2017/03/sp500_hist2014.gif)
A sample 500K portfolio bought at the beginning of 2018 would consist of:
500k in a 14yr bond ladder yielding ~3.1%
15.5k of Jan '19 LEAP options on XSP with a strike ~10% OTM which works out to be 31 contracts with a $300 strike purchase at ~$500 (SP500 was at ~$2700).
31 contracts at $300 represents exposure to 31*100*$300 = $930k for any movement above 10% in the underlying index for the year which represents about 186% leverage.
My backtesting to 1928 is showing stupid returns:
5yr  CAGR  12.46%; 80% confidence between 4 and 21% with a max of 29% and a min of 1.5%
10yr  CAGR  12.33%; 80% confidence between 7 and 17% with a max of 22% and a min of 4%
30yr  CAGR  12.26%; 80% confidence between 10 and 14.5% with a max of 16% and min of 9%
This strategy would be best in a IRA as the tax consequences on the bond proceeds would eat into returns. What am I missing here?

I'm curious why you're not also laddering the duration of the options? Wouldn't it make sense to be long options at a variety of expiries out to the longest dated call you can buy?
Interesting post.

Thanks for posting. This is not a bad strategy, especially for somebody who can't weather a decade long downturn.
When you backtest, if I understand correct, you are using current vix level and it's low historically. So that might be inflating your returns a bit. Still not a bad strategy.

That is a good point about the VIX being historically low right now. I checked some trades earlier this year on SPX options which I have traded for a while and it seems like LEAPS are priced off the interperprolation between the two realized volatility indices which correspond to their expiration date. i.e I bought some $300 strike SPX LEAPS on April 9th with 240 days till expiration. The 6 month vol index was 15.2% and the 12 month was 12%. The black scholes model:
http://www.cboe.com/framed/IVolframed.aspx?content=https%3a%2f%2fcboe.ivolatility.com%2fcalc%2findex.j%3fcontract%3d96E5FB367EDC4485A01002D7284CE7EC§ionName=SEC_TRADING_TOOLS&title=CBOE%20%20IVolatility%20Services (http://www.cboe.com/framed/IVolframed.aspx?content=https%3a%2f%2fcboe.ivolatility.com%2fcalc%2findex.j%3fcontract%3d96E5FB367EDC4485A01002D7284CE7EC§ionName=SEC_TRADING_TOOLS&title=CBOE%20%20IVolatility%20Services)
Says implied volatility was 13.2 based on what I paid. The price of a the 12 month LEAP option should correspond to the 12 month volatility index.
(https://qph.fs.quoracdn.net/mainqimga1f8ef78c1f847a1743d328fe03fcfb9)
8ish years of data can be downloaded here: https://us.spindices.com/indexfamily/riskindicators/realizedvolatility (https://us.spindices.com/indexfamily/riskindicators/realizedvolatility)
Playing with the black scholes model, it looks like a 365 day call option costs:
$4.5 @11% Vol.  back test 10yr CAGR  13.6%
$7.8 @14.4% Vol.  back test 10yr CAGR  8.2% (average 1 yr Vol. over time)
$13.4@20% Vol.  back test 10yr CAGR  4.8%
$18.5@25% Vol.  back test 10yr CAGR  3.5%
I think I need to figure out a way to backtest the strategy using implied volatility over time versus the return for the market that year. Laddering the LEAP options might be one way to take advantage of a more normalized volatility profile rather than getting lucky and buying when volatility is low.

I think one approach you could take is to think about in which scenarios this strategy could "blow up" or underperform, if ever?

Does your math imply that the average return over 100 years on a 12 month S&P 500 call option is 300%? It seems like it if you are getting 12% returns on the entire principal while risking only the 3.1% interest payments on the options. I haven't done the math but that just seems really unlikely that options could have such a high expected return. Wouldn't you get mind boggling numbers if you ignored the whole CD thing and just say bet 10% of your money annually on the options?

I think one approach you could take is to think about in which scenarios this strategy could "blow up" or underperform, if ever?
One scenario is a slow rising, low volatility market. For instance, if the market grinds forward 5% one year, 0% the next, 10% the year after, plus dividends, you are not getting any of that gain if you have 10% OTM calls.

Does your math imply that the average return over 100 years on a 12 month S&P 500 call option is 300%? It seems like it if you are getting 12% returns on the entire principal while risking only the 3.1% interest payments on the options. I haven't done the math but that just seems really unlikely that options could have such a high expected return. Wouldn't you get mind boggling numbers if you ignored the whole CD thing and just say bet 10% of your money annually on the options?
That's the way the math works. Take my example from this year. Buy the 31 $300 strike option contracts for $5. If the market goes up by 15% ($310) your contacts are worth 31k or a 6.2% gain on the 500k principal. The market goes up 20%, you make 14.9% on the principal. Historically the market exceeds 20% returns 25% of the time and 10% returns 45% of the time. The 12% (probably 89% with normalized volatility) is really lumpy but there is no downside risk to the principal other than inflation risk.

What is the ticker for this XSP product?

What is the ticker for this XSP product?
http://www.cboe.com/products/stockindexoptionsspxrutmsciftse/sp500indexoptions/minispxindexoptionsxsp (http://www.cboe.com/products/stockindexoptionsspxrutmsciftse/sp500indexoptions/minispxindexoptionsxsp)
XSP represents 1/10th of an SPX index future. In practice I have been buying SPX options and filling odd lots (19) with XSP options because SPX options have more liquidity.
http://www.cboe.com/products/stockindexoptionsspxrutmsciftse/sp500indexoptions (http://www.cboe.com/products/stockindexoptionsspxrutmsciftse/sp500indexoptions)
In Interactive Brokers, just search SPX or XSP under order ticket and select options.

I think one approach you could take is to think about in which scenarios this strategy could "blow up" or underperform, if ever?
One scenario is a slow rising, low volatility market. For instance, if the market grinds forward 5% one year, 0% the next, 10% the year after, plus dividends, you are not getting any of that gain if you have 10% OTM calls.
I would agree this strategy would underperform in your scenario. You can obviously adjust the % by which your call is OTM to compensate. At 10% OTM, you would expect the calls to expire worthless ~60% of the time. The caveat to buying calls closer to the money is the amount of leverage you can buy goes down. In back testing, the further out of the money the calls the more this strategy returns but the returns are even more lumpy. You can play with buying two different strike values say 5% and 12% otm or buy bull spreads, say buy the 5% OTM sell the 15% OTM capping your returns but allowing more leverage between 515%.
I think the worst case for this strategy is actually high volatility making the LEAP options more expensive thus decreasing the leverage. Think Jan 2009, the market returned 26% that year but this VIX started around 40ish. I have no idea what the 12 mo. realized volatility index was at. This is the VIX1Y:
(http://www.cboe.com/libimages/defaultsource/defaultcboelibrary/vix1y2.jpg?sfvrsn=7b844348_0)
The realized volatility index was probably a bit lower, maybe 30. SP500 was at 890 so to buy calls at 980 (+10%) Black Scholes says Jan 2010 leap options would have cost about $8.10. This means you could have only bought 19 XSP leaps. The market at expiration of those leaps was $1144. So you made (114.498.0)*100*19 = $31,160 or 6.2%.
If volatility had been at 14.4 (average) those calls would have cost $2.75 and you would have bought 56 calls. (114.498.0)*100*56 = $91,940 or 18.3%.
The ideal scenario is low volatility high returns:
Let's look at 2013. The market returned ~30%. 2 Jan 2013 12m Vol. was 12.96%. SP500 was ~1426 and ended 1/17/2014 at ~1848. 10% OTM leaps cost $3.65 so you bought 42 leaps. Gain was (184.8157)*100*42= $116,760 or 23%

Hi Ross,
How do you allocate the funds between taxable and taxsheltered accounts?
My guess, with the caveat that I know little about the US taxation system:
Sheltered accounts: 100% fixed income.
Taxable accounts: whatever is left of the fixed income pie, plus calls.

Hi Ross,
How do you allocate the funds between taxable and taxsheltered accounts?
My guess, with the caveat that I know little about the US taxation system:
Sheltered accounts: 100% fixed income.
Taxable accounts: whatever is left of the fixed income pie, plus calls.
That is correct. The SPX calls are taxed on a 60/40 long/short basis:
Under section 1256 of the Tax Code, profit and loss on transactions in certain exchangetraded options, including SPX, are entitled to be taxed at a rate equal to 60% longterm and 40% shortterm capital gain or loss, provided that the investor involved and the strategy employed satisfy the criteria of the Tax Code.