Is there a large counterexample?

True or False?

If $a,b$ and $c$ are roots to the equation $x^3+x^2 + x=1$, and let $S_n = a^n + b^n + c^n$, then there exists a positive integer $n$ such that $\text{sgn}(S_n) = \text{sgn}(S_{n+1}) = \text{sgn}(S_{n+2})$.

$$ Notation: $\text{sgn}(x) := \begin{cases} \begin{array} {l l } -1 & \text{ if }x<0 \\ 0 & \text{ if }x=0 \\ 1 & \text{ if }x>0 \\ \end{array} \end{cases}$ denotes the sign function.