He doesn't mean 40% of the population died. He means that actual deaths higher than expected/anticipated mortality based upon historic deathrates. I haven't looked at the underlying numbers recently but did so during the first two years of the Covid crisis. It was another way of measuring the impact of the pandemic. We know that each year about x% of Americans (or other demographic category) die. So if x+3% die in a given year, the mortality was 3% higher than expected. That might be due to chance, disease, pandemic, accident, war, malnutrition, drought, etc. I can't imagine a 40% rate, so I'd be a bit skeptical but would accept the statistical analysis of a legit agency.