In a recent Seeking Alpha article by Reel Ken, titled " Selling Puts - Investing Made Easy," he stated that a put writing strategy on the SPDR S&P 500 Trust ETF (SPY) can generate a better return than just investing in the SPY and re-investing the dividends. His analysis covered the period from January 2000 through July 2012, in which the market returned a total return of 1.78%, including re-invested dividends.
In my article titled "Simulating Market Price," I simulated the SPY using a geometric Brownian model and suggested that the model can be used to test trading strategies. This is the real pay off for developing models like these. That is, modeling can provide you with better insight than just back-testing historical data. With the use of this model, I am able to test this strategy over many potential SPY markets, and determine the conditions under which a put write strategy outperforms a fully invested approach. The simulated market that the model produces is a synthetic market that could possibly occur.
The strategy is tested by comparing an investment in SPY, with re- invested dividends, to a mechanical put writing strategy. I call the put writing strategy mechanical because puts are mechanically written monthly no matter what the underlying SPY price is at expiration. If the SPY is below the previous strike price, and the SPY is assigned, the position is sold and puts are written on the next trading day. If there is no assignment, and the put expires worthless, then puts are written the trading day following expiration, no matter what the underlying SPY price is. That is, there is no attempt to optimize.
The structure of the model is as follows:
1. Start with two account balances of $100,000.
2. In one account, buy as many whole shares as I can of the SPY on the first day of the model data and re-invest the quarterly dividend payments. (I call this the fully invested account balance.)
3. In the second account (which I call the put strategy account balance), I write a put the first trading day after the monthly expiration date of the contracts.
4. I determine the whole number of contracts that the put strategy account balance will buy at the current market price.
5. Monthly puts are written at or near the money on the trading day after expiration Friday. I round the market price up or down to determine the closest strike price to the money.
6. The put premium is calculated with the following linear formula and added to the put strategy account balance:
% Put Premium = .05 * number of days to expiration + .25.
7. On each succeeding expiration day, I determine if puts expires worthless or if there is an assignment.
8. If the puts expire worthless, I write another round of puts the trading day after expiration. The current put strategy account balance dictates how many contracts I can write.
9. If assigned, I debit the put strategy account balance by the number of contracts times the strike price and credit the put strategy account balance by the number of contracts times the current market price of the SPY. I sell all shares of SPY and write another round of puts the trading day after expiration. The current put strategy account balance dictates how many contracts I can write.
10. At the end of a year's worth of data I sell all the shares in the fully invested account, including the re-invested dividends, and calculate the gain or loss.
11. The put account balance at the end of the year is compared to the beginning of the year to calculate the gain or loss.
Results of the Simulation
The table below shows a small sample of 10 trial results that the model produced. It compares the return that was produced by the fully invested strategy to the put write strategy.
|Trial||Fully Invested||Put Write Strategy|
Notice that even with 10 trials, the average return over the 10 trials was very close for the fully invested strategy vs. the put write strategy. The difference can be seen when the fully invested strategy returns are very low, as was the case in the Reel Ken article. For low or negative market returns, the put write strategy appears to do better. The fully invested strategy produces a better return than the put write strategy when the market returns are high (I am arbitrarily defining high as 15% or more).
Conclusions and Potential Model Refinements
Model results show that the put write strategy beats a fully invested strategy for down, flat, or weak markets. This agrees with what Reel Ken found with historic data. Strong markets, however, favor a fully invested strategy. The model results are intuitive in that writing puts produces more income than dividends produce, but limit your upside.
I made no attempt to optimize the put write strategy. Many variations may exist that might improve the return. I have not simulated a put write strategy using weekly options. It would be very interesting to see if weekly options offered any advantage. Last, I have used a liner equation to calculate a put premium to account for time value differences. I would expect premiums to increase when the VIX spikes. A refinement to the modeling could involve adjusting the put premium to account for fear in the market.