Even on simple CSV data, Text::CSV_XS (not Text::CSV) can outperform plain perl (using split) given the right options and using the right methods. YMMV. Note, for the given data it most likely will not be faster

The main reason to start using Text::CSV_XS (or Text::CSV) even with simple CSV data is that moving onward when the data changes (yes, it will, eventually) or on other projects that look simple but are not.

For programmers that know both approaches, you are definitely right. No single modules (or tool or even language) is the best choice for every problem, but when in doubt, choose the option that is most versatile and best documented and that will also help you with likewise problems in the (near) future.

Even the simplest change, like changing the line ending from \n to \r\n will make you happy having made that choice as early as possible.