It is wrong.
The paper you have linked is a survey (economic historians are like other people and believe things that do not have clear evidence too, as someone who studied economic history I can give you a long list of subjects on which opinions without evidence are common...this is one of the most notorious), it does not say that it contributed half (there is no way to know this either, it is an anti-factual statement, there is research that says it contributed to the drop in imports...but this is against the backdrop of a massive drop that was probably at least 5-10x as large caused by banking), the quantum is extremely important here because you can say something is probably negative but also probably irrelevant (true in this case, the reason why this statement is said is because tariffs are negative ceterius paribus, so it is easy to say that they were negative but this ignores all other context...the irrationality about tariffs is exposed by almost all of the growth miracles in economic history occurring in countries with extremely high tariffs), and (finally) there is massive amounts of evidence that 99% of the cause was banking.
On the latter, this is knowable because you can point to failures of specific banks that coincided with the Depression getting worse in areas where those banks traded (in particular, the failure of Caldwell). This is a very different kind of evidence to the one for tariffs, in economic history terms the latter is shrug maybe (this kind of thing is not apparent to people who don't know how the sausage is made). This is why you have papers (like Eichengreen) that revolve around asking why SH is such an obsession for economists (usually not actual economic historians). Compare this to the number of papers on banking history of the period, on the failures of massive banks like Caldwell...there are very few on this because banking history is extremely unpopular and boring amongst economists because you can't use mathematical models that show how clever you are, macro is very popular but completely useless (again, most people don't know how the sausage is made).
There is no evidence that it contributed significantly. This is like your house being on fire, and saying that your house collapsed because you left the kitchen door open (and, again, to repeat: there is no evidence that tariffs are bad either...because almost every country that has experienced huge growth had tariffs in the past, there is a lot of evidence that tariffs/trade barriers are bad for economically uncompetitive countries i.e. the EU today, South America in the 50/60s, and Britain 20-70s but those two things are not separable, tariffs have a context).
Other comments are also mostly wrong. Issue wasn't unregulated banks in the GD either, most banks that failed were regulated. There is an argument for saying that state regulators were worse, that there was massive regulatory fragmentation (in the 20s, banks were regulated in a completely different way to today) but I am not clear why people would assume regulation automatically leads to less crises. Savings and Loans were also heavily regulated...still blew up. The issue is that heavy regulation usually causes massive concentration in the banking sectors (Canada and Australia are two examples) and this generally leads to a much lower frequency of banking crises but significantly greater severity. The assumption that regulators can just magically find this optimum is not logical (and is based on the theory that people who work at banks do not have an incentive to stop failures, this aspect was sold heavily after 2008 to support significantly more regulation...but it isn't accurate, for example Lehman's senior management lost 95% of their net worth, and ignores that regulators were overseeing the institutions that failed before too).