Perhaps this was the first semester that what had been known to a few people, or whispered about, became widely known. Perhaps this was the first time a digitized copy of the test bank got widely distributed -- maybe even because the publisher's standards for selling the test bank got loosened.

At a smaller but growing scale, over several semesters, this tactic of using the test bank might not lead to any striking anomalies in the distribution... just a gradual rise in the average, and fattening of the above-the-median tail. For a while, that might convince the professor he's doing better!

(Also, while we see Summer scores for comparison, he doesn't show the prior Spring or Fall. Often Summer students are a different pool, who in this case may not have been privy to the same campus folklore. Perhaps there was evidence of the same tactic earlier, or another professor with a different exam style taught previously.)

Lots of speculation, I know, and I would need more information to assign levels of culpability overall. Were students specifically warned off using test bank questions? Did those who obtained the questions lie to the publisher to acquire them? How similar were the circulated test bank questions to other official preparation materials? Etc.

It could also be interesting to see, on the makeup exam with presumably all-new but similarly-difficult questions, has the average mastery gone up? Perhaps memorizing the test bank questions does have some persistent beneficial effect for understanding the material.