My answer was that because SVM is a soft margin hyperplane optimization problem (not a maximum margin hyperplane). But he said that it is a simple question but u are not able to answer. it is because of hing loss. no more explanation.