Kernal estiamte of variance

[Richard Ramsawak]

Hi all

I am estimating a non-parametric model between two variables growth rate of cities and city size. I have been able to estimate mean standardized growth rate using Epanechnikov kernel using a bandwidth of 0.5, based on

lpoly StdGrowth2 lnScale2, bwidth(.5) ci noscatter reps(5000)

I am trying to estimate the kernel estimate of variance, using the same bandwidth, but I have not been able to find the relevant coding

I came across kdens, which generates a point estimate of variance, but the model does not run


kdens StdGrowth1, var(V) ci vce(bootstrap, reps(500))

But the code does not work

I keep getting an error


st_store(): 3200 conformability error
st_kdens(): - function returned error
<istmt>: - function returned error
r(3200);

end of do-file

Any help or guidance will be appreciated

Best

Richard