Let $f\colon Y\to X$ be a morphism of schemes. Assume $f$ is finitely presented, flat, with geometrically reduced fibers. Then Romagny has proved that the "functor of relative geometric connected components", that sends any $X$-scheme $T$ to the set of open subschemes $U\hookrightarrow Y\times_X T$ such that $U_t$ is a connected component of $Y_t$ for every geometric point $t$ of $T$, is representable by an algebraic space Z, étale and finitely presented over $X$.
Assume now that every connected component of a fiber of $f$ is geometrically connected. Is it true that $Z$ is a scheme ? The étale map $Z\to X$ will then have split fibers (i.e., $Z_x$ is the direct sum of finitely many copies of $\mathrm{Spec}\; \kappa(x)$ for all $x\in X$), and I have the impression that the should imply that $Z$ is a scheme, but my intuition can be wrong.