While creating a percentile to give context to net positioning is a step forward, it’s still a flawed approach for one reason: you have to account for markets growing and shrinking over time. In the above example, what if when commercials had their huge net short position of 555,715 contracts the open interest in corn futures was 2,088,225 contracts? Now, the market is smaller since the open interest is 1,707,744. So the current net short position of 426,928 is “bigger” relative to the total size of the market than it would have been when open interest was +20% higher.
The solution is simple. Rather than calculating percentiles based on net position numbers, I now calculate percentiles based on net position numbers as a percentage of open interest. Using the exact same data from above, here’s the data adjusted for OI:
Most Bullish Position: -12,363 contracts on 11/1/2013 (when OI was 1,902,900) = -0.65%
Most Bearish Position: -555,715 contracts on 6/17/2016 (when OI was 2,088,225) = -26.61%
Current Position: -426,928 contracts on 3/10/2017 (when OI was 1,707,744) = -24.99%
Using the above three numbers, the 5-year percentile is now 15%. This is much less than the previous reading of 24%. Why? Because open interest in corn has decreased and the current net short position of -426,928 is bigger relative to the size of the market than it would have been a few years ago. Let’s dive into two graphs that show the differences between these two calculations. First, the previous method: