Smile! It increases your face value

Sammanfattning: This thesis examines some of the multiple variations of the previously established Rules of Thumb; which are used in attempting to explain implied volatility sur- faces. Here, these Rules are extensively tested on the Swedish stock market index (OMXS30) using rolling window analysis and linear stepwise regressions with for- ward selection. Data is collected on options at four different maturities and nine different levels of strike prices for ten years (2512 trading days), resulting in a lin- early interpolated volatility surface consisting of 36 points for each trading day. The results support previous research on other stock indices in that the Square Root of Time Rule is the most accurate according to both in-sample information criteria and out-of-sample measures of fit. According to the results, the most ac- curate regression specification for one-day-ahead prediction is the Square Root of Time Rule with, in contrast to previous research, an observation window of two days and including explanatory variables up to twelve powers.