This issue extends to every arbitrage I have ever thought about. Suppose that there are two classes of stock. Class A trades at a 10% discount to class B. Should you hedge one-to-one or one-to-0.9?
The answer depends upon your time horizon and the risks of the trade. For instance, suppose that the fund portfolios are quite volatile. If fund A and fund B doubles, what do you think will happen to the relative discounts? If they stay the same and you are on a hedge based upon the value of the underlying value of the portfolio, you'll get creamed.
(Example: Suppose that both portfolios have $10 worth of stock and A trades at a 10% discount while B trades at a 0% discount. A is trading at $9 and B trades at $10, so you buy 1 shares of A and short 1 share of B to be hedged based upon the value of the underlying portfolio. The portfolio value jumps to $20. If A stays at a 10% discount, it will go to $18. If B stays at 0% discount (as it will if it's an ETF), it's value goes to $20. Since you are on a one-to-one hedge, you lose a dollar.)
However, if the discount narrows, a one-to-one hedge will be more appropriate.
In general, I use a market value hedge BUT I adjust the hedge as the discount narrows and widens. (So in the the example above, I would buy one share of A and sell short 0.9 shares of B. As the discount widens or narrows, I adjust the hedge.) Essentially, I'm taking the risk the correlation between the discount and the value of the underlying portfolio is negative. That is, if the % discount narrows as the stock goes up and lowers as the stock price goes down, I'll need to buy more shares of B at a higher price and sell more shares of B at a lower price. My experience has been the opposite, so I stick to my market value hedge.
