An advantage of having EconOverflow SE would be that it would hopefully then become non-contentious that Economics SE could be for those who study, teach and apply economics and econometrics, with low-level questions accepted (subject to conditions such as effort shown). In other words, a structure similar to MathOverflow SE and Mathematics SE.
In response to the question in the comment by @EconJohn, I offer the following examples relating to a topic I happen to have studied:
Question suitable for Economics SE (but towards the high end) In the context of estimating the recreational value of a national park using the zonal travel cost method, what methods might be used to allow for heteroscedasticity in estimating the trip-generating function, and what are their respective pros and cons?
Question suitable for EconOverflow I am trying to estimate the recreational value of a national park and have zonal travel cost data for 15 zones. With so few data points, conventional methods for addressing heteroscedasticity in estimating the trip-generating function (eg weighted least squares with weightings estimated from the data, heteroscedasticity-consistent standard errors) seem unlikely to work very well. An alternative approach proposed in Christensen & Price (1982) looks promising as its assumptions are independent of the number of zones. Which method might be expected to produce results with the least bias?