Negative margins are valid in css and understanding their (compliant) behaviour is mainly based on the box model and margin collapsing. While certain scenarios are more complex, a lot of common mistakes can be avoided after studying the spec.

For instance, rendering of your sample code is guided by the css spec as described in calculating heights and margins for absolutely positioned non-replaced elements.

If I were to make a graphical representation, I'd probably go with something like this (not to scale):

The margin box lost 8px on the top, however this does not affect the content & padding boxes. Because your element is absolutely positioned, moving the element 8px up does not cause any further disturbance to the layout; with static in-flow content that's not always the case.

Bonus:

Still need convincing that reading specs is the way to go (as opposed to articles like this)? I see you're trying to vertically center the element, so why do you have to set margin-top:-8px; and not margin-top:-50%;?

Well, vertical centering in CSS is harder than it should be. When setting even top or bottom margins in %, the value is calculated as a percentage always relative to the width of the containing block. This is rather a common pitfall and the quirk is rarely described outside of w3 docos