Comparing coefficients of linear " Stochastic Frontier Production and Cost Functions" in R I am using the function frontier::sfa() in R to obtain the Stochastic Frontier Production and Cost Function as followed:
stbl1 <- x <- frontier::sfa(
                           formula = nostemsLog ~ 1 + qmdLog,
                           data    = dat1
                           )

I am doing this for two datasets (dat1 and dat2) and want to compare if the coefficients of the resulting estimates stbl1 and stbl2
> coef(stbl1)
(Intercept)      qmdLog     sigmaSq       gamma 
  4.7566418  -1.3577710   0.5378605   0.9326808 

particularly Intercept and qmdLog are significantly different. I am unsure how to proceed, as the concept of Stochastic Frontier Functions is still a bit foreign to me.
How can I do these test in R?
Background: I am fitting self thinning lines for different species and want to compare these.
Edit 1
One reproducible example based on the package:
> library(frontier)
> data(front41Data)
> dat1 <- front41Data[1:30,]
> dat2 <- front41Data[30:60,]
> x1 <- sfa(log(output) ~ log(capital), data=dat1)
> x2 <- sfa(log(output) ~ log(capital), data=dat2)
Warning message:
In sfa(log(output) ~ log(capital), data = dat2) :
  the parameter 'gamma' is close to the boundary of the parameter space [0,1]: this can cause convergence problems and can negatively affect the validity and reliability of statistical tests and might be caused by model misspecification
> x1

Call:
sfa(formula = log(output) ~ log(capital), data = dat1)

Maximum likelihood estimates
 (Intercept)  log(capital)       sigmaSq         gamma  
      2.8646        0.2642        0.4364        0.8243  
> x2

Call:
sfa(formula = log(output) ~ log(capital), data = dat2)

Maximum likelihood estimates
 (Intercept)  log(capital)       sigmaSq         gamma  
      2.7035        0.4550        0.9736        0.9972  
> 

The warning above is only an artefact of the example and I don't have it in my actual data.
I now want to see if the coefficients coef(x1)[1] is significantly different from coef(x2)[1] and if coef(x1)[2] is significantly different from coef(x2)[2]. 
I am only interested in the slope and the aspect. One paper where they used this approach is Charru et al. (2011), although it is not really clear which arguments they used for the sfa() function.
 A: The specific "package" ‘frontier’ version 0.996-10 for R, that is mentioned in the paper you linked appears to estimate a "half-normal production" stochastic frontier, meaning that the specification assumes that the composite error term is comprised of a zero-mean symmetric normal random variable minus a half-normal r.v. (and to do justice to the authors of the paper, they describe clearly the theoretical framework, in their Section 3). 
In "production" SFA models, the predicted dependent variable estimates the maximum, "frontier" value of the actual dependent variable. Experience has shown that in many cases the obtained coefficient estimates differ little from the corresponding OLS ones (meaning that OLS also is close to estimate the "frontier values"). The value added of SFA lies mainly in the decomposition of the residual variance. Since for the iterative maximum likelihood estimation algorithm we need starting values, we obtain these by first running OLS. It is always useful to review the OLS estimation and compare results as regards coefficient estimates.
This being maximum likelihood estimation, asymptotic properties of the various statistical tests hold. So, if you had run two OLS regressions, how would you test whether the coefficients that interest you are significantly different or not between the two regressors? You can do the same here.
