As far as I know it doesn't actually say that. The comparison is between countries, not before and after reforms. The paradox is supposedly that countries with higher equality have higher gender differences. But it isn't really a paradox since correlation isn't causation.