Reserves are required because things go wrong (stated rigorously: unforeseen, in-scope efforts). If estimates are not met, or work is discovered that was not part of the bid, the money to comply comes from somewhere. Perhaps, if an engineer misses a requirement, it can be recovered with some uncompensated effort. But uncompensated overtime is rarely enough to cover the additional work uncovered in most projects.
Most of my familiarity with reserves is with Non Recurring Costs, but any design to cost activity must also contain reserves, for the same reasons as below. Any budgeting process (even design tolerance budgeting, for example) must contain reserves to cover the risk inherent in the estimating process. Or your rocket may not get off the pad.
Even with the best efforts, some tasks will go over, and presumably, some will come in under. But don't count on anyone coming under. Most people will see to it they use their entire allocated budget to reduce their personal risk on the task assigned.
A reserve is absolutely necessary, but hidden, abused, and denied at every turn. People are always finding better (more expensive) ways to do things, and encountering problems that require more effort. The PM can feel like they are the only one in the company actually trying to meet the budget. If the PM is getting no help in this regard, then the reserve is their only hope of being successful, budget wise.
Strictly speaking, a reserve is set aside to cover unexpected contingencies. This is controlled by the program manager, although this may be allocated to each cost account manager in the Earned Value system, since they are mini-program managers.
Sometimes the reserve is established by simply planning using 90% of the budget, holding back 10% as the reserve. In order for this to be effective, each manager has to attempt to do the work at 90% of the budget. This usually does not work. Most people will not change anything when given a 10% budget challenge. So why would they come in under? But it is a common practice, and you will encounter it.
A more complex, but more effective way to generate a reserve budget, is to estimate the work going as planned, and also include estimates for the risks involved with probabilities and impacts to others. Estimates then look like this, using a software task as an example:
Incorporating the probabilities of the example above yields an estimate of 115 hrs. The base estimate is 100 hrs. The engineer shoudl work to that. 15 hours is placed in the reserve pool. Remember, since this is probabilistic, the reserve may be spent on other tasks.
There is also an IMPACT to flag to others to consider this risk during their estimates.
The example above is very simple, but it should be clear how to set this up. Sum all the hours, and all the probabilities. Then you can generate a minimum value if everything goes as planned (a statistical impossibility), a weighted probability value (most likely), and a worst case value (also impossible). The risk budgets are then quantitative. The larger the number of tasks in your risk pool, the better this works. You do not know how the reserve will be allocated, but you know what people were thinking when it was bid. Once again, managers have to try to work to their budgets. And don't forget to throw in some hours for "unknown unknowns".
This allows the reserve to be set in a much more transparent manner, and the effect of something not panning out is defined and can be allocated for. So if something goes wrong, theoretically, the PM already has a budget to cover that.
There are plenty of studies and papers on reserves. Some attempt to eliminate the concept entirely. But let's get to the title of the article.
You absolutely need one, and most likely have one, but actually using one is a bitch.
Starting at the beginning, a reserve can never be identified as such in a proposal - and the word can never be used with the customer during negotiations (see section on customer language). It implies you have wiggle room in your estimates, and it will be dog meat during negotiations. It is very difficult to even discuss cost of risk, even internally among peers. (What risks? There are no risks!)
What happened in the old days was that the managers padded their estimates for risk, and the program mangers lopped off a percentage to create a reserve, which brought the quote down to about where a nominal estimate should have been. But it was never discussed as such. Things have evolved since then.
To combat the reluctance to discuss risk and its cost, some customers now invoke formal Risk assessments that are are contract requirement. (The Project Management Institute even has a Standard Practice for Risk). This requires discussion of the risks. However, this is in direct conflict with normal customer interface, so is usually strained. I have not been impressed with the results of these exercises, but at least it is a step in the right direction.
During the execution of the program, if managers know about the magnitude of the reserve, it may become something that they may depend upon - "don't worry, go ahead, we have a reserve. We will still meet our original estimate". If all the managers do this, then there is no reserve, and the reserve concept is defeated. You overrun. The managers have to try to meet the minimum estimate for this to work.
A major misuse of the reserve occurs when the program manager uses the risk budget for other purposes than coping with approaches that do not work out. I have seen PMs use this funding to support customer appeasement, by agreeing to scope increases with no corresponding cost increase. This makes the customer happier, but unfortunately diverts this budget from its original intent. From a probability standpoint, it insures the program will go over budget. But the customer will be happy!
The best way to handle this is to establish a fund for little victories for the customer. Remember, the way your customer is graded may be on how much extra stuff he got without a cost increase. Set up something to give them so they get their bonus points, or they may become very unhappy, and aggressive in their pursuit of contract performance wins.
So the Reserve is a budget item you cannot admit exists, but everyone knows it is there, and everyone would really like a piece of it - if not all of it. People may even plan their work based on getting this unknown quantity. And your boss certainly knows about it, and may have plans for it, as well as your corporate office, and even the customer. They may have already spent it before you encounter your first problem.
But handling issues like this is why they pay you the big Yankee dollar.