I trade both. In my case:
- stocks pair trading: Smaller positions, higher trading frequency (holding periods of about 1 week), smaller profit per trade, and more diversity (about 10-20 pairs open at any one time). I generate the pairs and divergence signals quantitatively, with very little discretion (I only verify pairs are in the same industry, and no earth shattering news came out).
- futures calendar spreads: Larger positions (contract nominal size), longer holding period (about 1 month), larger profit or loss per trade, and more concentration (1-2 spreads open at any one time). I look at the story (planting, feeding, etc) and spread charts, and my selection of the spreads and entry-exit points are fairly discretionary.
My primary focus is quantitative trading, so I prefer (my) pairs trading over (my) futures calendar spreads. They are both profitable over long term, but the calendar spreads have sharper drawdowns.