Password policies should be evidence-based.
Is the system stand-alone, or internet linked? Long, complex, frequently changing passwords aren't that necessary if you have 24/7 multiple security guards with computer nodes in locked windowless rooms in locked buildings behind razor-wire fences with intrusion detection and hidden claymore mines.
How critical is the information on the systems?
If the data or system isn't a critical success factor for your organization, then strict security rules aren't necessary. Of course that begs the question of why you have a computer system in the first place.
Does the value of the information on the system decrease with time? If so, set up your security so that it's not impossible to break into, just not possible to break into before the information is useless to the theives.
What are the costs to the organization if the system is compromised? If it means that terrorists can spot times to blow up a nuclear power station without getting caught, then you need very secure password policies (along with other authentication measures). If it means that someone can walk off with patient health information, then it could cost the organization $100 to $25,000 to $50,000 or more. Then you have to balance the annual costs of resetting passwords with the cost if the information is improperly accessed.
Put it in dollars and cents to the policy makers. At least that way you have a REASON for the policy, and not some knee-jerk idea pulled out of the fat man's buttocks.