When I read about it obviously I was furious. After all, the commutative property is such a basic understanding of mathematics.
But this article written in English talks about it and explains the situation, and I sort of understood the rationale:
http://d.hatena.ne.jp/takehikom/20131117/1384646400
"There are 5 cars. How many tires do you have altogether?" is the word problem which is presented first. Sharing a tip that each car has 4 tires, the pupils are expected to write "4 x 5 = 20." If someone wrote "5 x 4", then the expression would represent "4 five-wheeled vehicles" or "4 houses each of which has 5 cars". They can understand visually that either situation is not what is shown in the original problem.
To the best of my knowledge, the oldest example is found in the prototype of Elementary School Teaching Guide for the Japanese Course of Study, Mathematics, published in 1951 [URL 5]. It is reported that more than a few pupils replied the incorrect expression in calculating the number of pencils using multiplication, where the writer inferred from their reactions that they lined the two numbers up and sandwiched in the times sign without thinking deeply, and that they were incapable of organizing the situation for a valid choice of operation. The teacher told the pupils that the expression did not fit in with the problem because it would yield the number of persons. Here B-4 was adopted.
Mod Edit:
The answer is circled as correct but the formula is circled as incorrect.