It can be negative.

It represents the proportion of variance in $y$ explained by $x$.

It is equal to the slope of the regression line.

It is always equal to $1$ in real data.