Niels asks: "Is it really important to have a SP estimate before the first Sprint starts?" I think he means an estimate of the initial team velocity before the first Sprint Planning Meeting.
Answer: The short answer to his immediate question: Yes, it is important.
Why? It makes it a tighter meeting (SPM, Part1), and therefore the business stakeholders are more likely to show up regularly. This is good because they understand the Demo far better if they came to the Sprint Planning Meeting. They can see, often enough, that the key problem is them: they (a) did not give the PO enough time, (b) did not let the PO talk to their people enough, and (c) don't really understand what they want.
But let's look at the bigger question inside Neils' question. What is really important?
Of course this is a question that has puzzled mankind and Socrates for many years.
In terms of Scrum, I would say that delivering Business Value to the customer is most important. And then other things are important to the degree they support that. Among them, that the Team should have a better life.
Having an initial velocity estimate is not the most important thing in the world. But it helps, and the cost is low.
People can also misunderstand the initial estimate, and you should discuss that also.
How does the initial estimate help? The initial plan for the release is somewhat more realistic. The initial Sprint Planning Meeting is less chaotic. The business stakeholders will find Part 1 of the meeting somewhat more useful (less disorganized). The work taken into the Sprint will be closer to what the Team can really do. The Team is somewhat more likely to feel positive traction and 'success' after the first Sprint. All these things help.
But let us not mistake these positives for our real goal. Our real goal is to satisfy the customers with high business value.