Estimates are a vital part to ensure that each project is done on time and within budget.  Over the years, I have seen various patterns of estimators out there -- here are the top 5 ones that you don't want to have on any project followed by their remedy.  Here are the processes that I believe some people use that as a project manager you need to recognize and correct the behavior:

• The "Padder" - "No matter what I believe my estimates are, I'm going to add a fudge factor to it!".  This is what you would hear from this pattern.  This kind of estimator will take risks and uncertainty into their own hands instead of getting the project manager involved.  People who do this usually don't trust their historical estimates.   What's worse is that the project manager if they don't see this may also end up adding to the estimates, thus "double-dipping" -- causing estimates that are way overstated.  Remedy:  Make sure that this kind of estimator has the opportunity to identify risks, questions and other things that cause the uncertainty.  Also, show how past estimates from this person were padded and make sure that they trust their estimates.
• The "Utopian" - "If everything goes right, my estimates will be correct".  This kind of estimator looks at every task as half full, and isn't willing to accept that there are risks.  Remedy: Murphy's law is always in effect, and it is the job of the project manager to uncover every rock and bring to light any risks.
• The "Pessimist" - Polar opposite of the Utopia, and worse than the "Padder".  This person will determine every single scenario that could occur on every task and take the very worst scenario. Remedy:  It's the project manager's job to remove risks and other obstacles during a project, so they need to make sure that these kinds of estimators bring those issues up with plans to remove those issues prior to the actual tasks.
• The "Back Filler" or "Make My Date" - "We have pre-determined dates anyway for every project right?  So, I'm going to make sure my estimates hit that date regardless if its accurate.  If I was wrong in my estimates, I'll just work more time in order to hit the date."  Though it is tempting by both project manager and team members to make the project schedule hit the date, this is the wrong approach in doing so.  Any time I have seen this pattern, I have seen those people work a lot of overtime and they usually still miss their dates because they dramatically cut their estimates because of an unrealistic date. Remedy: No matter how much pressure you get from project stakeholders to hit a date, educate them on the importance of understanding the scope of the work first, then adding resources or cutting scope in order to hit a particular date.  Make sure that the team members understand the importance of their estimates through this process.
• The "Generalist" - "I'll give you a rough idea though high-level estimates, but I can't guarantee any more details than that."  This kind of pattern indicates that the estimator either doesn't trust the process, doesn't know how to estimate, or is the results of poor requirements and design in understanding the scope of the project.  Remedy: Make sure that each person that is required to estimate truly understands the scope of the project.  Through a Work Breakdown Structure (WBS), show how the estimator can break down their tasks in order to better understand what they are going to do on the project.  With better details of how they will do the work, will come better estimates.  Also, make sure that they don't feel penalized later on the project as discoveries are made, that the schedule will adjust as necessary for changes.  Through this, they will trust the process.