I don't believe you're calculating vacancy rates the same way the article is. You're talking about single-unit vacancy over time, the article is referring to citywide vacancy over volume. Depending on how that number if calculated those could be wildly divergent values.