rdplot error

[Adrij Chakraborty]

I am trying to run an age-based local linear RDD with optimal bandwidth, using individual level cross-sectional data. Y1 is my outcome variable. rvv1 is my running variable. mage_imp is my monthly age variable (which is the uncentered version of the running variable). My specification is the following:

  rdrobust Y1 rvv1 if imp_m==0, all h(`hopt') b(`bopt') kernel(triangular) vce(cluster mage_imp) covs(illness_dummy* department_dummy* wave_dummy2)

I use all the specified dummies for fixed effects.

If I drop observations exactly at the cutoff (i.e., drop if rvv1 == 0), I get normal standard errors (nothing unusual). But if I increase the donut-hole size by even 1 month (i.e., drop if in range(rvv1, -1, 1)), then I get odd standard errors like 2.5e-14 (unusually low) for the conventional estimate. I think I can still ignore it since I focus on the Robust estimates and standard errors for my analysis.

When I wish to plot the RD plot (keeping a donut hole of 1 month), using

 rdplot Y1 rvv1 if imp_m==0, p(1) h(`hopt') kernel(triangular) vce(cluster mage_imp) covs(illness_dummy* de
> partment_dummy* wave_dummy2) masspoints(adjust) covs_eval(mean) ci(95) shade graph_options(xtitle(Dist. from 5 years (mont
> hs)) ytitle(Utilization Probability) xlabel(-`hopt' 0 `hopt', format(%9.2f) angle(90)) scheme(st1color) graphregion(color(
> white))) ///
> name(YC1, replace)

I get the following error:

 rangen():  3351  argument has missing values
                 <istmt>:     -  function returned error
r(3351);

Could anyone tell me if:

1. The odd SEs and the missing argument errors are related?
2. Why do I see this error during the plotting of my RDD, and what I can do (or check) to solve it?

Thanks in advance!