Let $X$ and $Y$ be non-negative r.v. defined on the same probability space $(\Omega,\mathcal{F},\mathbb{P})$ such that
$$\mathbb{P}(Y>y)\leq\frac1y \int_{Y\geq y}^{} X d\mathbb{P} $$
Show that for $p>1$,
$$\mathbb{E}[Y^p]\leq\Big(\frac{p}{p-1}\Big)^p \mathbb{E}[X^p]$$
Suggestion:
Recall that for a non-negative random variable $X$,
$$||X||_p^p=\mathbb{E}[X^p]=\int_{0}^{\infty}px^{p-1}\mathbb{P}(X>x)$$
and Holder's inequality states that for $p,q$ such that $\frac1p + \frac1q = 1$,
$$\mathbb{E}[ZW]\leq||Z||_p||W||_q=\mathbb{E}[Z^p]^{1/p}\mathbb{E}[W^q]^{1/q}$$
Big hints:
Response to comment:
$$\int_0^\infty p y^{p-2} \int_\Omega X 1_{Y \ge y} \,dP \, dy = p \int_\Omega X \int_0^Y y^{p-2} \, dy \, dP$$