Think opportunity cost. PhD you are looking at 4-5 years, masters 2 max. When my sister got into the PhD track at Stanford, they told her that if she had a job offer, she should probably take the offer, because odds are that her earnings after PhD would never make up for the lost earnings during the period. Her program can be characterized as computational biology, probably one of the fastest growing fields (imho finance is in a secular decline, with both pay and number of jobs to shrink). Of course you could start a PhD at as school with a quant program, take a bunch of classes and cop out of PhD, but you won't make yourself very many friends in faculty. Further, with a PhD, you price yourself out of lower level entry jobs, which means that your supply of available jobs will be smaller (of course the supply of candidates is also smaller, but is worth trade of?) You might see that the demand for PhD's will decrease by the time you are done.
While research typically requires a PhD, this is not a hard rule. I know someone with a BComm doing research at a small quant fund, though he did teach himself stochastic calculus and is fluent in a host of programming languages.
If you have something you want to study and thirst the knowledge, I would suggest a PhD. If you are just trying to get into industry, go for masters. If you go into PhD with the sole goal of getting a job in industry, it will probably be rough, because faculty will sniff that out pretty quick, and not all will be very receptive to that.
