A characterization of the gamma distribution in terms of conditional moment (Q686355)

Extending a previous result by \textit{S. Osaki} and \textit{X. Li} [IEEE Trans. Reliab. R-37, No. 4, 379-382 (1988; Zbl 0659.62017)] the author proves the following characterization of the gamma distribution: A real-valued random variable \(X\geq 0\) with probability density \(f\in C^ 1((0,\infty))\) follows a gamma distribution (parameters \(\lambda\geq 0,1)\) if and only if \[ E(X^ k| X\geq y)=\left[\sum^{k- 1}_{i=0}{\lambda+k-1\choose i} i! y^{k-i}\right]r(y)+{\lambda+k- 1\choose k}k! \] holds true for some \(k\in\mathbb{N}\) and all \(y\geq 0\), \(r(y):=f(y)/\int^ \infty_ yf(x)dx\).