Suppose you have a choice between two jobs in two cities.
Job A is in City Z, Job B is in City Y; Job A pays 100,000 a year, Job B pays 50,000 a year, but offsetting this is that things are twice as expensive in City Z as in City Y.
Which job is better?
The answer depends on what year it is. In the year 1900, they're exactly the same. In the year 2010 (putting taxes aside for the moment) Job A is definitively better, because even though things are twice as expensive in City Z, there's nothing stopping you from ordering your luxury goods shipped from City Y.
If you are a company hoping to produce goods to sell, which city looks better?
Well, you can open up shop in City Y, pay your employees half as much, and ship your goods to City Z. City Y is much better.
Unless you are providing local services, of course, in which case City Z is more attractive, because your 5% markup on your goods becomes more substantial.
Companies and their employees have fundamentally conflicting interests in terms of location. This is a specific case of the broader problem that is the coexistence of high wages and unemployment.
Other cases include minimum wages and what I call Lottery Jobs - jobs which normally have absurdly low pay, but which have rare individuals making absurd salaries attracting large numbers of other people into the field. These are typically the kinds of jobs armchair economists refer to as "sexy" - which is to say, attractive on a basis apart from the salary. Owners of bars, fashion models, actors/actresses, game developer/designer (your flash game may be fun to design and play, but isn't likely to net you the fortunes of Peter Molyneux). These are jobs at which one is likely to barely be able to afford to live, much less make the fabled fortunes. (Even the success of a game like Dwarf Fortress doesn't net its creator massive fortunes; he makes less than I do, although I suspect he has lots of room to grow yet.)