I think it's really hard to tell. There are so many factors to consider. I'm inclined to agree that on average, a PhD (even in CS/Stats) isn't going to help someone make more money -- they'd probably make a bit less, especially over 5-10 years, than someone without. But at the very least, there will be a huge variance on that average.
These comparisons can be quite subtle. A PhD, and what one does during a PhD, could help someone who would not stand out from their undergraduate record/CV get a job that typically hires bachelors recipients. He or she may make the same as bachelors degree holders in that position, but they have a job they wouldn't have otherwise gotten. Factoring these sorts of scenarios together makes it hard to say, even on average, what monetary gain a PhD will bring. I know of someone who would have really struggled to get jobs with his CV after undergrad, but now has a PhD and a good publication record, and would really stand out on the job market.
In general, the connections can open a surprising number of doors in the long run. And after a few years, good quants can make $800,000/year. But I wouldn't encourage a PhD to someone who wouldn't at least enjoy the process.