There are a number of reasons why only a small percent of options are exercised. For calls:
1) If the call is ITM and still has time premium remaining, exercising it throws away the time premium. Therefore it makes more sense to sell-to-close the option. If coupled with someone who is short the call and is buying to close, the contract ceases to exist and Open Interest declines.
2) It is capital intensive to exercise. Suppose XYZ is $202 near expiration and I own 10 calls ITM worth $2 with a strike price of $200. If I exercise, I need $200k ($100k on margin) to accomplish this. All you need to buy long options is the cost of them. Sell to close the calls and no additional capital is needed.
3) Apropos to the year this question was asked, every trade involves a commission. If you buy and then sell to close the call. it's two transactions. If you buy the call, exercise and then sell the underlying, it's three commissions plus additional bid/ask slippage (unless you trade at one of the few brokers who does not charge a fee for assignment and exercise).
4) On the flip side, consider the writer of the call. Suppose it's a covered call near expiration, the short call is slightly ITM and the investor does not want to lose his stock. Either he buys to close his short call or he rolls it out to a later week/month (still BTC). If coupled with a STC transaction, this too results in a reduction in Open Interest and the contract ceases to exist.
5) Short sellers don't want the expiration risk of a position in the underlying. For example, you sold a $45/50 bearish put vertical spread for $2. You maximum risk is $3 because of the protective long $45 put. On expiration day the stock is $47 and you have a $1 loss. If you do nothing, the $45 put expires, the $50 put will be assigned and on Monday morning you're long the stock with significantly more downside risk. BTC the short put to avoid this risk.
The general theme here is that it's almost always better to close positions than to exercise.