What makes you think Capers Jones' 2017 study didn't take testing into account? His study covered the total life cycle cost of software projects.

There is no programming language that obviates the need for thorough testing. No programming language can guarantee defect-free software.

I've been a software engineer for over 30 years. I've written tons of code in Fortran, C/C++, Java, C#, Python, Go, Smalltalk, etc. and I've never found statically typed languages to be better than dynamically typed in any fashion.

Robert Smallshire conducted a study that showed only 2 percent of software defects may be attributed to dynamic typing. This is not significant and certainly not worth abandoning the many benefits of dynamically typed languages.

Half of the top 20 programming languages in the world are dynamically typed, and for good reasons.