This 2% number that keeps being thrown around is based on dividing the number of confirmed deaths (which is likely to be relatively accurate) by the number of confirmed infections (which is almost certainly a massive underestimation). It's basically meaningless.