As an IT Contractor in London for the last 20 years, I have had to negotiate salary many many times. To summarise my experience, and how I approach the whole process:

1. I would normally mention a ball-park figure to any agency or client trying to recruit me. If they know the client can only pay less than my minimum number, then it's time to agree not to go further with the process.




2. However, I don't reveal previous rate/salary or expectations in detail. Ball-park only.




3. Once interviewing begins, try to avoid talking about money at all, until you have got them 'hooked'. Interviews should be about 'Do you fit the job', not about 'How much do you want?'. I would normally only talk cash in a second+ interview, if possible.




4. If they ask how much you want, DO NOT ANSWER directly. You must fudge it, or answer with a question: I normally ask them back how much they are prepared to offer. If you answer, then it can only go down. If they answer, then it can only go up. So yes, the first person to mention a number loses.




5. If they come back with a number that's too low, sometimes the best thing to do is not to respond, except perhaps to say 'Hmmm', and look a bit downcast. I've had instant improvements with this technique! Or, ask if there is any way they can squeeze the budget to bring it up, as you were hoping for a bit more. That works too. If that doesn't work, you should outline politely why you think you have the attributes that would attract a better pay rate. Don't be pushy, but make it clear that you would not be happy working at that rate.




6. Remember, they are spending a lot of time and effort selecting candidates, and dealing with (usually wretched) agencies. If they finally get to someone good, they will be prepared to pay, and the person you are going to work for will be pushing on your behalf.

Finally, a true story. I went for interviews at a big bank. Great job, nice people. At the 3rd interview, my future boss asked how much I wanted for daily rate. I asked what he could pay. He asked what I wanted...and so on, as we both knew the 'game'. Finally, I said 'Tell you what. Why don't you go to your budget people, see what your best offer is, and call me with that? If it's OK, I won't haggle or mess around, because I want the job, and I'd be great in this position.'
So that's what he did, and the next day he called me with a number much much higher than I had been hoping for. So I danced a little jig, then quietly thanked him and said I'd be starting next week.

The basic theory behind "the first to give a number loses," from the job seeker's perspective, is that one might be "leaving money on the table" by naming a salary request. This viewpoint has a deeper root, however, in "positional bargaining," in which the two parties' interests are directly opposed and the relationship is adversarial.

A strict "positional bargaining" strategy, however, is usually not in the best interest of either party to the negotiation. Instead of worrying about who names the first number, do your homework. Determine in advance what your interests are and what you would be willing to accept. The overall package (including benefits, location, type of work, opportunities for advancement, etc.) consists of far more than just salary. Also consider what is typical for hiring someone of your skill level for the type of work you are doing for the type of company in that geographic area. I know that is a lot to consider, but if you do your research, you should be able to arrive at some conclusions.

At that point, you can be prepared for the salary negotiation without worrying about who names the first number. If you are at the point of salary negotiation, the company already wants to hire you. Get on the same side of the metaphorical table -- you are not adversaries! You are pursuing a mutually beneficial relationship. Talk about what is fair instead of simply making salary demands. If you can do this effectively, you can show that you are easy to work with while still maintaining your interests. Further, you make the discussion about the issues, not the people, so no one gets offended.

While I find it aggravating that some companies ask for salary history (which is often irrelevant to what is fair, given what I just stated), if you have to give it, do so. It really does not change the negotiation if you get on the same side of the table. If the employer is unwilling to negotiate in this manner, you may wish to reconsider whether you want to work for that employer anyway.

I realize this answer may be vague, but I'm trying to make the point that, with negotiation based on interests rather than positions, naming numbers becomes far less relevant.

There's over 4 million hits on Google for "negotiation theory first offer". They provide differing viewpoints and a definitive answer is hard to find, as there are many variables influencing the outcome.

To name three:

1) A strong argument for making the first offer is the anchoring effect.

In situations of great ambiguity and uncertainty first offers have a strong anchoring effect—they exert a strong pull throughout the rest of the negotiation. - Harvard Business School

What that means is that when you make the first offer, the other party tends to forget his or her aspirations, becoming overly influenced by your first offer. The offer shapes the rest of the negotiation.

Even Dilbert know this: 1, 2

2) Negotiating salary is typically a win-win type of negotiation. Both parties can and will get what they want, or else no deal will be made. But they can be win-lose if one of the parties is under outside pressure to accept terms that he/she does not find agreeable. If this is the case for the job applicant, then it may not be wise to offer first.

People who are extroverted do worse in a win-lose negotiations. They tend to be agreeable and to be swayed by the opponent's first offer. - Wyche.com

3) A situation in which making the first offer is not to your advantage:

When the other side has much more information than you do about the item to be negotiated or about the relevant market or industry. - Harvard Business School

This is typically the case in negotiating salary. The employer often, but not always, knows how much they can pay.

So what I suggest you do:

• Try the exact same Google search
  


• Make an inventory of the factors involved in your particular case


• Determine if the factor works for or against you making the first offer (+1/-1)


• Give the factor a weight (what importance do you attribute to the factor?)


• Multiply these last two and add up the products. The outcome will be <0 or >0

This is essentially a weighted averages method of decision making. Then fit your decision to the real-life circumstances. Maybe the employer will offer first, and maybe it will be an offer you're more than happy with.

Not necessarily. I think you can shoot for the stars and ask for more than even you think you deserve (within reason.. maybe 10 or 20k more than you would accept). In that case, you can wind up winning even when you've named first.

It's also possible that a company will offer a number you can't accept and that they can't go above for budget reasons, in which case you both lose.

Special cases:

1. If you need to list your salary history you haven't necessarily
   lost. If you can justify a step up from your last position, a
   company may be willing to consider it. I would say (having done some
   recruiting) that you should try your best not to provide this
   information. It is rarely disqualifying if you don't provide it
   (unless not providing it prevents you from even completing an online
   application form).




2. You can give wide ball-park ranges to a company to see if you're in
   the right neighborhood of what they can pay. Based on job
   description alone you should be able to guess what sort of level a
   position is. The number of times you'll be completely wrong about
   the level of a job having seen the description will be far fewer
   than the number of times you're right.



3. I would say it's actually a great strategy to shoot high and then accept a little lower. It makes you seem easy to work with, yes, but more importantly it makes you seem like you know what you want and what you're worth.

You can't be too concerned about wasting time on an interview that turns out to be for a job that's way below your range. It won't happen often.

More importantly though, don't accept less than you want and count on earning more after reviews/promotions. It almost never works out well unless you have a formal review written into a contract with an agreed-upon raise if the review is favorable.

Bold answer

Is this true?

No.

Evidence

Anecdotal - two interviews, several months back.

In one, I didn't really want to work there and added a big, let me say outrageous, premium to my usual expectations (just to get them off my back without declining). They said it was too much, and I walked away. Then they got back to me twice, with better offers, ending up above what I would usually have asked.

In the other interview I told them my expectations - reasonable this time - and they came right out and said that I would be worth to them more than what I said and hiked the offer up from what I asked.

What does losing mean?

In this context, what does it mean, to lose? Leaving money on the table as an employee? Spending more than you might have had to as an employer?

Then winning means knowing, you didn't leave money on the table; you didn't spend too much. And that is something nobody can know.

It doesn't matter, who mentions the first number, if you do negotiate this way

everybody loses.

Everybody will leave the negotiation with this nagging question: "Could I have done better?"

So how can I win?

In the first interview I was willing - actually intended to - walk away. Willingness to step back from a deal puts you in a better position when negotiating, because you obviously can't be pressured as much.

On the other hand there are employers willing to pay an employee's valeue and not just the least amount they can possibly negotiate and still hire you. Winning is easy here.

The pattern though, in both cases is this:

Not playing the game by the commonly accepted rules.

Which are to set two boundaries and step by step find a common place in between.

Instead play a game I call "being okay".

• Find out where you stand and how you value your time. That is your price tag.


• 
  Be okay with that evaluation


• Learn if your potential employer respects that evaluation or if maybe they value even higher than that


• 
  Be okay with their reaction (there's no reason to be upset if they don't value you in the same way)


• If they respect your evaluation, accept, otherwise decline


• Be okay with the outcome

This is what I do and what I am happy with. Your results may vary, but I highly doubt it, if you are genuinely okay in each step.

The statement "the first to give a number loses" is fundamentally wrong.

The reality is "the one who doesn't know the value or isn't willing to walk away 'loses'."

Rather than trying to nickle and dime the potential employer during salary negotiations, you need to know ahead of time exactly what it is you want. Then you need to be prepared to walk out and not look back if you don't get that amount. You also need to be prepared to be happy with getting that amount, even if they say yes with zero negotiation.

If you are desperate, or just have to have this particular job for whatever reason, then you lose. If you don't have a very real sense of your worth, you lose.

However, if you understand that there are a lot of companies out there and are capable / willing to spend the time to find the diamond, then you will win. This absolutely requires you to know what the market rates are for what you do. If you have unrealistic expectations - $1m/year for recent college graduate - then you will be 'educated' along the way.

What about large employers who have the power to absolutely require
all candidates to reveal detailed salary history?

This doesn't matter. Yes, companies use that as a starting point. However, you aren't forced to accept their offer. If you were previously making $10/hour and now want $30/hour - and that's in line with the job market - then you can stand your ground on what you want. It's just a starting point to give them an estimate of what they think you'll be willing to take.

How do you avoid wasting everyone's time by being completely out of
range of what the company is willing to pay?

I ask the recruiter very early in the process what the pay range is. If it isn't at least close to what I want then I tell them where they need to be before I would even consider talking further. If they aren't willing to tell me, I decline to interview.

Isn't it possible to "lose" more by seeming difficult to work with
than to simply ask for a generous salary right at the beginning of the
discussion (when the recruiter/hiring manager brings it up, of
course)?

Only if you are desperate. The key here is to force the issue early on. You bring it up before walking in the door for the interview. For example, you submit an application, they call you to schedule an interview, before setting the time you ask what the pay range is. If they don't answer, don't waste your time. If they say something nebulous like "market rate" - ask them what that means to them. If the range is roughly what you want, even if they are 10% lower, then go interview. They'll ask what you want at that time, tell them.

Salary negotiation really isn't hard. The key is just to know what it is you want before you even submit an application. Everything that follows is just theater.

Actually, I've read that the reverse is true. The first offer sets a baseline and basically a framework to work within. Unless your expectations are way outside the initial range that the employer has in mind, you are better off making the first bid. Even if you are higher and the employer says, "That's a bit higher than what we have in mind for this position" you can counter with "Well what would be more in range?" They will respond with a higher number than what their initial number would have been and you can counter with a higher number.

Of course if you bid far lower and they bite, you're leaving a lot of money on the table, but you've also met your desired amount so in the end you still get what you wanted out of the negotiation.

Edit for references
I wish I could recall where I read that...I've ripped through a number of books on negotiation and sales over the past year, and at best it would be a wild guess which one of them I picked this up from. But I can tell you from personal experience (over many contract negotiations as an independent consultant) that I've rarely had to take a rate that I couldn't tolerate. And I always lead with my rate when the question comes up.

Of course some of my success also comes from applying other strategies. The biggest one for me is being able to walk away from a bad deal. Just the act of saying no thanks increases your bargaining power sufficiently. I've had several callbacks where they were willing to match my requested rate. Of course contract rates and salaries are different scenarios altogether. In a contract, the agent is usually trying to maximize their profit margins, for direct hire, the manager usually has a set budget range for a req. I wouldn't suggest no thanks as a bargaining strategy for direct hires because it will usually be a dead end.

Is this true?

It creates a baseline for negotiation — Gaining a substantial (two-digit-percent) increase on your salary is unlikely.

Your prospective employer will think in terms of "how much more should I give this person so they come and work here?" and not "how much is this person worth to me?".

What about large employers who have the power to absolutely require all candidates to reveal detailed salary history?

They don't have that power. They can only ask. You can answer "I will move for x$" if they don't accept that, say that you will not disclose the sum, but you will move for x$. If they drop you, they don't want you.

How do you avoid wasting everyone's time by being completely out of range of what the company is willing to pay?

Ask them what they want to pay, for example. There's no reason why they shouldn't disclose their range (and in fact, they do it all the time to recruiters!).

Isn't it possible to "lose" more by seeming difficult to work with than to simply ask for a generous salary right at the beginning of the discussion (when the recruiter/hiring manager brings it up, of course)?

No. At the end of the day, someone from recruitment will bargain with you, but someone from the specific department you will join will sponsor you. In the end, if they really want you, they'll go up within their allotted range.

I would say that the first person to mention a number draws the battlefield line. The other person then gets to choose whether to engage, and if so, whether they will accept the line as drawn or push back.

Think of a poker hand. Neither player knows the other player's hand, but they do know their own, so they get information about the other player's hand by betting (and watching the other player's bets). In such a situation, the player acting first must give information before getting any, by opening the pot with a bet. Then the other player gets three choices; call, raise, or fold. They also get to make this decision based on the information from your bet.

In this analogy, a "call" would be to accept your terms; the other party thinks you've bid in an equitable or even disadvantageous way for yourself and takes your offer. A "raise" would be a pushback on your terms; he thinks the situation is disadvantageous for him, but that he can improve on that position by continuing to play the "hand". A "fold" is a refusal to participate; the battle lines, as drawn, are so heavily in your favor that the safest play for the other player is to cut their losses and try again with the next hand. In the real-world equivalent, this is "Well thanks for your time, we'll let you know".

Now, while the player acting first doesn't have much information, he also gets to set the tone of the hand, and if he knows from experience or research how the other player(s) are likely to react to various bets, he can choose the one most advantageous for him in the situation presented by his own hand. Back in the real world, you can intentionally lowball your salary request in the initial interview, giving you a greater chance of getting the job, then later, when you've made yourself indispensable to the company through your knowledge and contributions, you have more bargaining power when angling for a raise. If you think the company is dying to have you at any cost, you can ask high, setting the bar at your advantage for further haggling. If you think your position and the company's are roughly equal, where you want the job but there are others, and the company wants you but there are others, then ask for the average salary for your experience and industry (which requires you to know what that figure is!

People who say that you shouldn't name a number first are presumably thinking of a scenario where the employer is prepared to pay, say, $100,000, and then you ask for $50,000 and they immediately agree and you've lost the potential higher pay.

But there are other possible scenarios that work against you. You refuse to give a number. They look at your resume and say to themselves, "Someone with these qualifications will probably demand $100,000, and we just can't afford that. No point wasting our time interviewing him." If you in fact were prepared to accept $50,000, you might have just finessed yourself out of a job.

Realistically, if you aren't going to take a job for less than $100,000 and the most the company is prepared to offer is $50,000, there's no point even scheduling an interview. You're just wasting both of your time.

If the company has no idea what the current market rate for this job is, by naming a number first you may establish an expectation. If when you walk in they're saying to themselves, "I wonder what people like this get paid these days. $50,000? $100,000?" If you say you expect $100,000, they might sigh and say oh well, I guess it's the upper end of what we were expecting. If they name a number first, they might well offer the low end. There's some psychology here: someone can get a "reasonable number" in their heads and be unwilling to move far from that, even if there was no reasonable basis for the "reasonable number".

And while it's no doubt fun to think that I might walk into a job interview and the company offers me ten times what I was expecting, realistically, that's not likely to happen.

I think in real life, when someone applies for a professional job, he probably has more control over the final salary negotiated than the employer does. If the applicant is presently employed, he's rarely going to accept a pay cut to take a new job. He's likely to be happy with a modest increase. So if he's presently making, say, $50k, he'll probably ask for $55k or so and be happy to accept $52k. The range between his asking price and his minimum is pretty small. From the employer's point of view, unless they're a very small company or this is a very high-paying job, what they pay this person is a small percentage of their total budget. If they were expecting to pay $50k and a candidate comes along who seems highly qualified but he demands $55k, they might well decide to just stretch a little and do it.

When I'm buying a product, I get very suspicious of a company that refuses to tell me their price. When I see an ad that says "call for pricing", I never call, because I take it for granted that if they had a good price, they would list it in the ad. If they want me to call, that must mean that their price is high, and they're hoping that if I call they can talk me into it with a slick sales pitch.

So my conclusion: I don't volunteer a salary number. But if someone asks, I give them a number.