In “Pygmalion in Management,” J. Sterling Livingston writes for HBR:

Managers’ high expectations must pass the test of reality before they can be translated into performance. To become self-fulfilling prophecies, expectations must be made of sterner stuff than the power of positive thinking or generalized confidence in one’s subordinates—helpful as these concepts may be for some other purposes. Subordinates will not be motivated to reach high levels of productivity unless they consider the boss’s high expectations realistic and achievable. If they are encouraged to strive for unattainable goals, they eventually give up trying and settle for results that are lower than they are capable of achieving. The experience of a large electrical manufacturing company demonstrates this; the company discovered that production actually declined if production quotas were set too high, because the workers simply stopped trying to meet them. In other words, the practice of “dangling the carrot just beyond the donkey’s reach,” endorsed by many managers, is not a good motivational device.

It’s important to match expectations up with evidence, and to also communicate this connection. If there’s no reason to believe that a stretch goal can be achieved (or is even meant to be), then it won’t be. That’s the fundamental tension in OKRs, the goal of which is to draw out results that can multiply a team member’s impact. (There are some other challenges worth noting with OKRs.)

Without this sense of evidence, there’s no grounding an expectation in reality. Deadlines become dreadlines. People feel like they’re being set up to fail, so they just don’t bother.