[R-sig-Geo] Standard heteroscedascity versus heteroscedascity in spatial econometrics

[Hermann Pythagore DONFOUET]

Dear Sir,
I'm working in spatial econometrics and I read some papers of Kelejian and Prucha (2007, 2010) about integrating heteroscedasticity in spatial econometrics. In spatial data, most often spatial units differ in terms of their shape and size. Thus, the hypothesis of constant innovations is unrealistic in the class of SARAR models (spatial effects in the dependent variable and error term).
But, I don't really understand the main difference between heteroscedascity in spatial econometrics and heteroscedascity in standard econometrics (a-spatial models). Please, do you knowm the main difference?
Thanks in advance and regards
Hermann

	[[alternative HTML version deleted]]