1: \begin{abstract}

2: We consider the problem of inferring the total causal effect of a single

3: continuous variable intervention on a (response) variable of interest. We propose a

4: certain marginal integration regression technique for a very general class

5: of potentially nonlinear structural equation models (SEMs) with known

6: structure, or at least known superset of adjustment 

7: variables: we call the procedure \emph{S-mint} regression. We easily

8:   derive that it achieves the convergence rate as for nonparametric regression: for example, single variable intervention effects can be estimated with convergence rate $n^{-2/5}$ assuming smoothness with twice differentiable functions.  

9: Our result can also be seen as a major robustness property with respect 

10: to model misspecification which goes much beyond the notion of double

11: robustness. Furthermore, when the structure of the SEM is 

12: not known, we can estimate (the equivalence class of) the directed acyclic graph

13: corresponding to the SEM, and then proceed by using \emph{S-mint} based on

14: these estimates. 

15: We empirically compare the \emph{S-mint} regression method 

16: with more classical approaches and argue that the former is indeed more

17: robust, more reliable and substantially simpler. 

18: \end{abstract}

19: