he is suggesting it as an explanation for retirement contributing to premature death in men more than in women, not the difference in life expectancy in general.
Men, as a hypothetical, might take up drinking and fishing, leading to an increase in mortality caused by liver failure, cancer, what have you.
Women, again as a hypothetical, might take up other activities that don't increase mortality.