I'm currently reading Paul Wilmott's excellent book on option pricing. Near the beginning, he constructs a risk-free portfolio using an option, and a short on the underlying to hedge the risk. I'm specifically interested in European options.
A no-arbitrage argument follows:
If this portfolio earns more than the risk free rate: borrow money at the risk-free rate, buy the portfolio, and make money off the arbitrage.
Conversely: short the portfolio, invest money in a risk-free instrument, and again make money off the arbitrage.
I've scoured the internet, but couldn't find an explanation for the second argument, which I have a hard time grasping! By shorting the portfolio, we short an option, and "short a short", meaning we go long on the stock.
So, when we short the portfolio, we might even have to spend additional money, if shorting the option didn't give enough money to buy the stock.
This segment focuses on the binomial model, so I've tried separating this to 3 cases:
When in both the up and down state the option is worth more than 0. In this case, the arbitrage relies on buying the amount of stock that can be had by exercising the option. I have a hard time finding arguments to why in this case the option should be worth more than the stock at the period before expiration.
When in both the up and down state the option is worth 0. I understand this case, the option is worth 0 at the turn before expiration, and the hedging is a degenerate case (longing 0 stocks).
When in the up state the option is worth > 0, and in the down state the option is worth = 0. Like in case 1, I can't find a good argument.
As you can see, I'm out of answers. I don't even understand why a risk-less portfolio must earn the risk-free rate. Anyone has a clue?