1: \begin{abstract}

2: In the past decades, weak convergence theory for stochastic processes

3: has become a standard

4: tool for analyzing the asymptotic properties of various statistics.

5: Routinely, weak convergence is considered in the space of bounded

6: functions equipped with the supremum metric. However, there are cases

7: when weak convergence in those spaces fails to hold.

8: Examples include empirical copula and tail dependence processes and residual

9: empirical processes in linear regression models in case the underlying

10: distributions lack a certain degree of smoothness. To resolve the

11: issue, a new metric for locally bounded functions is introduced and the

12: corresponding weak convergence theory is developed. Convergence with

13: respect to the new metric is related to epi- and hypo-convergence and is

14: weaker than uniform convergence. Still, for continuous limits, it is

15: equivalent to locally uniform convergence, whereas under mild side

16: conditions, it implies $L^p$ convergence. For the examples mentioned

17: above, weak convergence with respect to the new metric is established

18: in situations where it does not occur with respect to the supremum distance.

19: The results are applied to obtain asymptotic properties of resampling

20: procedures and goodness-of-fit tests.\looseness=-1

21: \end{abstract}