Several methods currently exist for extracting some of the information embedded in option prices. To date, most of these methods have focused on the second or higher moments of the underlying asset’s instantaneous return or value at expiration. In contrast to these methods, I present a method here which exploits transactions level options data to help forecast the first moment of lead returns in the underlying asset. This is accomplished by constructing a signal from what I call the “put-call-parity bias” of option quotes. Although put-call-parity intrinsically ties together the price of a put and call with the same strike and expiration, in the presence of transaction costs it is possible to observe bullish and bearish pressure based on the relative values of the bid and ask quotes for the put, call, index and risk free bond. Instead of working with these bid and ask prices directly though, I map these price into a set of 4 implied volatility values to form individual signals, and aggregate such signals using local polynomial regression; these aggregate signals are then used to forecast market returns. For the case of S&P 500 index options, these aggregate signals are shown to be related to recent and contemporaneous market performance, and more importantly appear to possess informational content for forecasting lead index returns.
Back