Casualties and Rules

The latest numbers show that the US air campaign against ISIS is killing more civilians under Trump than it did under Obama.  Should we blame Trump?  Sure, but perhaps not entirely.  I think several factors may be at work:

• urban warfare is just brutal.  No way around it.  The offensive to take Mosul has reminded folks of the line from Vietnam: "we had to destroy it in order to save it."
• US troops are now deployed in Syria, which means air strikes to protect them.  The Special Operations Forces training/assisting the Syrian allies are few in number and thus vulnerable.  The Marines and others deployed to provide artillery and other support are also relatively few in number and vulnerable.  So, when various forces--ISIS, Iranians, Assad's forces, etc--get close, air strikes happen.  And I am guessing the rules governing airstrikes to protect US troops at risk are probably different from the rules governing attacks on ISIS bases, etc.
• Trump.  Mattis and others have insisted that the rules haven't changed.  Maybe not, but rules are always interpreted.  One can bend the rules for a friend (as in the case of interpreting caveats in Afghanistan back in the day).  One may strictly interpret the rules (zero tolerance or whatever) if one is being watched very carefully by a superior (a principal, an overseer) especially when being caught has consequences. In Afghanistan, various Dutch officers liberally interpreted the rules because they knew there would be little risk of punishment, for example (again, see the book).  So, the Trump effects here are:
• Trump has signaled via his statements that he does not care about civilian casualties.
• Trump has delegated pretty much everything to the military--there is probably no concern that the National Security Council folks are watching, unlike during the Obama administration.
• Trump himself breaks all the rules, so as a role model, he inspires .... less strict observance of the rules.

War is constantly a gray area--if the rules say that a strike should not happen if it puts 15 people at risk (just an example as the rules of engagement are classified), then does the person calling in the airstrike say that there are 14 or 16?  Lots of estimates with big +/- uncertainty.  So, it is hard to judge.  But the trends do seem to be pretty significant.  Lots of things are in play, but I'd bet that the US armed forces a wee bit less careful now than when they were concerned they were being watched closely.  It is just basic human behavior (and principal-agent dynamics).