On D. Haegele's approach to the Bessis-Moussa-Villani conjecture