I want to prove that if $\alpha >0, a\in\mathbb{R}$ then $$\mathbb{E}|X|^\alpha<\infty \Longleftrightarrow \mathbb{E}|X-a|^\alpha<\infty$$ $\Longleftarrow$ is trivial taking $a=0$. However I am having trouble with $\Longrightarrow$. One of the strategies I have tried is proving for $X=\textbf{1}_A$ which guarantees the result for simple functions, then for positive ones and then for any real $X$. However I don't think I can use this here since I need to use the linearity of the integral and the absolute value and the power don't allow me to do this.
I have also tried Jensen's inequality since $|\cdot|^\alpha$ is convex if $\alpha \geq 1$, but this has been a failure as well. Finally, I attempted to write the random variable $X$ as an integral of an indicator function as done here, but again the power is bothering me.
I am sure there is a very easy way to prove this but I am very stuck. Can someone help me?