1: \begin{abstract}

2: Nonparametric data envelopment analysis (DEA) estimators have been

3: widely applied in analysis of productive efficiency. Typically they are

4: defined in terms of convex-hulls of the observed combinations of

5: $\mathrm{inputs}\times\mathrm{outputs}$ in a sample of enterprises. The

6: shape of the convex-hull relies on a hypothesis on the shape of the

7: technology, defined as the boundary of the set of technically

8: attainable points in the $\mathrm{inputs}\times\mathrm{outputs}$ space.

9: So far, only the statistical properties of the smallest convex

10: polyhedron enveloping the data points has been considered which

11: corresponds to a situation where the technology presents variable

12: returns-to-scale (VRS). This paper analyzes the case where the most

13: common constant returns-to-scale (CRS) hypothesis is assumed. Here the

14: DEA is defined as the smallest conical-hull with vertex at the origin

15: enveloping the cloud of observed points. In this paper we determine the

16: asymptotic properties of this estimator, showing that the rate of

17: convergence is better than for the VRS estimator. We derive also its

18: asymptotic sampling distribution with a practical way to simulate it.

19: This allows to define a bias-corrected estimator and to build

20: confidence intervals for the frontier. We compare in a simulated

21: example the bias-corrected estimator with the original conical-hull

22: estimator and show its superiority in terms of median squared error.

23: \end{abstract}