Counterfactual predictions

[kartechbabu]

Hello,
Is there any way I can make counterfactual predictions with CI after estimating the causal parameters? If so, what method is available to make these predictions?

Thanks,
Karthik

[SvenKlaassen]

Hello,

currently there is no method to make counterfactual predictions (especially with CI since this would include confidence intervals for all the nuisance elements e.g. in the PLR model this would require to also quantify the statistical uncertainty of g(X)) implemented.
Generally I would guess that these predictions should be very noisy in most settings.

I guess the closest part would be something like Heterogeneous treatment effects.

[kartechbabu]

Thank you so much, @SvenKlaassen, for your prompt response!