Specifically you're a company and you're trying to hire people to fill roles. If they are are required to be at the company you have to pay them enough that they can live near you.
You set up a series of constraints on hiring:
1) Compensate[0] enough so that someone from outside can enter the local economy effectively.
2) Compensate enough to convince someone who is already in the local economy to change from where they work to working for you.
3) Compensate enough that someone working for you won't go work somewhere else.
Within that system to have to operate a business model where the value you capture in the market has to cover not only your operating costs (which include compensation for the employees creating that value) but allows for you to fund maintenance, depreciation, and some R&D efforts for new product development.
If you can't find a solution set that works you have to improve the business model (generally add revenue or cut costs). Not really rocket science of course.
I keep watching to see if some large presence in SF ends up moving to the east bay or even Reno as a way of changing the cost parameters. If you look at job changing as brownian motion you can watch an environment parameters start to change which results in people (the particles in this case) start changing where they are likely to be.
Personally I would love some quantitative data collection that helped map out the weights of the various factors. It would help predict the next 10 years.
[0] Compensation is also a complex thing, more than just dollars per week/month/year there are things like free food, bus rides, and other services in addition to environmental plusses and minuses and work demands.