Chesterton's fence
This is a named principle rather than a measured effect. The logic is sound, and the documented cases keep bearing it out. Trust the direction, and treat the boundaries as open.
The writer G. K. Chesterton offered a principle about reform using the image of a fence across a road. Imagine you come upon a fence that seems to serve no purpose, standing inconveniently in the middle of nowhere. The impatient reformer says, I see no use for this, let us clear it away. Chesterton argued the opposite. The fact that you cannot see the reason for the fence is a reason to pause rather than a reason to remove it. Someone built that fence, deliberately, presumably for a reason. Until you understand why it was put there, you are not qualified to take it down, because the reason may still hold, and in removing what you do not understand, you may cause exactly the problem the fence was quietly preventing. Only once you understand why it exists are you in a position to judge whether it can safely go.
Before you remove or change something that seems pointless, first understand why it exists, because the reason may still be doing quiet, invisible work. Rules, processes, and structures that look useless or inconvenient were usually put there deliberately, often to prevent a problem that is no longer visible precisely because the structure has been preventing it. The confident move of sweeping away what you do not understand is how well-meaning reformers cause damage, reviving the very problem the old rule was solving. The point is to earn the right to remove something by first understanding what it was for, rather than to never change anything. If you cannot explain why it exists, you are not yet ready to take it down.
One says do not remove what you cannot explain. The other says that in a system you do not fully understand, removal is the higher-confidence move. The difference is what you know about the thing. Via negativa applies when you can already see it is actively hurting, as Semmelweis could with the unwashed hands. Chesterton applies when it merely looks pointless and you have not yet found what it is preventing. Removing a known harm and removing an unexplained rule are different acts, and only the first earns the confidence.
Sunk cost says what you have already put in is gone and should not steer what you do next. Chesterton says what was put there before holds a reason you have not recovered yet. They are not about the same past. One is money, time and effort already spent, which cannot be recovered and should carry no weight. The other is the information encoded in an earlier decision, which can be recovered and should be, before you act. Ask "would I start this today" about the cost, and "why was this here" about the rule.
Source: G. K. Chesterton, The Thing, 1929.
James C. Scott, 1998
On what goes wrong when planners remove or override the things they do not understand.
Draw your own card. It does not take long, and it rewards taking your time.