A study aims to estimate the total population $Y$. Using an auxiliary variable $X$ known for the entire population, the ratio estimator is $\hat{Y}_R = \frac{\bar{y}}{\bar{x}} X$. When is the ratio estimator more efficient than the simple expansion estimator $\hat{Y} = N\bar{y}$?