Re-examining the 'more people less erosion' hypothesis: special case or wider trend?