Correction to: Positivity (2020) 24:1399–1417 https://doi.org/10.1007/s11117-020-00737-x

Dutta et al. [3] have recently published a correction of the well-known result of Dutta and Chandra [2, Theorem 3.5]. Using a counterexample, they proved that Dutta and Chandra’s result is not correct and then provided a true statement of it (see [3, Theorem 2.1]) under the assumption that the contingent cone to the feasible set at the considered point is convex.

As the conclusion of [2, Theorem 3.5] was used in the proof of [1, Theorem 10], specifically for transitioning from

$$\begin{aligned} \underset{\eta \in \partial ^{*}h\left( \overline{x}\right) }{\sup } \left\langle \eta ,d\right\rangle \ge 0,\ \ \ \text {\ for all }d\in T\left( E,\overline{x}\right) \end{aligned}$$

to

$$\begin{aligned} 0\in \overline{co}\left( \partial ^{*}h\left( \overline{x}\right) +\left[ T\left( E,\overline{x}\right) ^{\circ }\right] \right) , \end{aligned}$$

we have to correct our results by adding the missing convexity of the contingent cone \(T\left( E,\overline{x}\right) \) to the statement of [1, Theorem 10] and the subsequent outcomes. This additional condition is crucial for the validity of our results.