Considering the CFTC's policy on position limits, there are actually two questions that need to be answered.
The first is whether it is true that speculative positions affect prices. There are good economic reasons to reject this notion. A futures position is a commitment to deliver a commodity by the short side and a commitment to take delivery of that commodity by the long side. Those commitments are distinct from demand or supply of the product. This is because the long side can immediately sell the commodity received to someone who wants to consume it. Likewise the short side can buy the commodity immediately before delivering it per their contract. For the long side, these constitute transfers of the commodity as opposed to demands for the purpose of consuming it. Likewise, generally the short side satisfies its delivery requirement via a transfer rather than through engaging in its actual production. Examinations of the effect that undertaking positions has on prices have consistently found no relationship between position changes and subsequent price changes. This is precisely the result that the above economic reasoning predicts. That is, unless long positions actually shift commodity demanded or short positions actually shift commodity supplied, then trading in futures should have no effect on pricing fundamentals.
But suppose one concludes that empirical tests are flawed either because the necessary data are unavailable or because the test procedure lacks power. Then it is reasonable to conclude that the above economic rationale is unproven. But that conclusion does not support a policy of setting position limits. This is because lack of evidence for the conclusion that positions affect prices implies the lack of a metric for setting appropriate limits. This is an important point. Absent an ability to say by how much prices are affected by speculative pressure, one also lacks the ability to state the size of the limit necessary to remedy the problem. This is true even if one is absolutely certain that speculative positions affect prices. Lacking a measure of the effect, one can only guess as to the correct size of the limit.
This need to guess implies that setting limits comes with substantial risks. Setting the limit too high, one risks not addressing the problems created by speculators. Setting them too low, one risks diminishing the ability of hedgers to transfer unwanted risks and, therefore, bearing risks that exceed their tolerance levels. Hence, getting the policy wrong has consequences for the economy, consequences that are impossible to anticipate and likely difficult even ex post to measure owing to the disparate nature of its impact on market participants.
*I just received this text on position limits from an "anonymous economist"