"Sell Mortimer! Sell!"
The variance premium in developed equity markets is a well established phenomenon. It persists. Anyone with an equity portfolio should take advantage of it. And you really should be trading options to do so.
But even after accepting this, you still need to decide which particular options to sell. SPY alone has thousands to choose from. Two recent papers help with this decision.
The first is "Understanding and Trading the Term Structure of Volatility" by Campasano and Linn. They look at the dynamics of implied volatility, and show it depends both on the slope of the term structure and also the maturity of the options. It is the maturity aspect that is relevant to us. They find that the returns to short one month straddles is about three and a half times that of short six month straddles.
The other relevant paper is "Which Index Options Should You Sell" by Israelov and Tummala of AQR. They look at returns for a given unit of risk. The most interesting unit in my view was their “stress test”. This evaluates the expected portfolio losses during extreme equity index return scenarios, i.e. "crashes".
They find that short-dated options with strikes just below the index price are the best in terms of profit when normalized by units of stress scenario losses. The reason they postulate is that option buyers typically purchase slightly out-of-the-money puts for insurance. So these options, which are most valuable to insurance buyers, have the highest insurance premium.
But for those of us who are already convinced of the existence of the variance premium, the specific reasons are of less interest than the conclusions: sell short-dated options just below the at-the-money.
You think they are too risky? Well the actual evidence suggests otherwise. And maybe ask yourself why the variance premium exists at all…