The problem of decreasing new candidates is clearly a real issue. As this chart details, the amount of new admins being promoted is much lower than it used to be. As recently as January 2008, there were more admins promoted in a month (36) than there were all of last year (28). This year got off to a good start, as we saw 14 successful RfA in the first three months (as opposed to five successful RfA in the first three months of 2012). The chart cited above also demonstrates that the total number of RfA (including unsuccessful) has been declining on a yearly basis, starting in 2008. In 2007, there were 920 RfA. Last year, there were 92. That is an astounding 90% drop!
What is causing the drastic decreases detailed above? Well, I'd imagine many of you have witnessed a scenario in which a prospective candidate is approached on their talk page by someone who offers to nominate them for adminship. The offer is declined, perhaps vehemently, as the prospective candidate does not want to go through the tortures of an RfA. It is no secret that the RfA process can be hard on a candidate. While one could argue that the rigors of the week of an RfA could be seen as preparation for adminship, more often than not, an RfA is unsuccessful and the candidate may not have much to show for his trouble. The high level of scrutiny at an RfA is often a deterrent for qualified candidates and understandably so. I recently wrote an essay titled "Admins are people too"; perhaps a similar essay for admin candidates is in order.
Different editors have different standards for RfA candidates. Arguing the importance of tangible things (like edit count and the candidate's promoted content) causes enough problems. Things can get really interesting when !voters begin exploring the subjective issue of a candidate's personality and if he or she has a demeanor suitable for an admin. Obviously, the standards are higher than they used to be. In 2005, a candidate for adminship began his self-nom statement with the following: "I've been here roughly nine months now, accumulating just about right at 1000 edits." He passed. Nowadays, that would lead to either a SNOW or NOTNOW, probably in a matter of hours. The standards have obviously gotten tougher and that trend could very well continue.
While the decreasing number of admins is cause for concern, it is not necessarily creating major problems just yet. We do have admin backlogs from time to time, and they can be frustrating, but the situation is not yet unworkable. Eventually, however, we recognize that it may be. Therefore, much time and energy has been spent trying to resolve the problem of lessening activity.
Recent major efforts include the attempted RfA reform in 2011 and a lengthy RfC from earlier this year. Obviously, the former has done little, at least noticeably, to improve the situation. The jury is still out on the RfC, but other than a WikiProject for nominators of RfA candidates, it has not yet resulted in anything substantial. The WikiProject has potential, but will it help to increase activity? Only time will tell.
A plethora of new or recycled proposals have been provided over time. It has been argued that we should "lower the bar", allowing candidates to pass with a lower percentage of support than currently required (we know that consensus is not just numbers, but numbers are very important at RfA). There are those who would be more willing to support a candidate if taking away the tools when necessary was easier to do. Others think that unbundling (giving out some of the administrator privileges to non–administrators) might help to prevent major admin backlogs in the future, but enough editors are opposed to this concept in general that it remains unlikely and cannot be relied upon as a solution.
At this point, I do not believe that prolonged discussions attempting to solve the problem are really the way to go. If a perfect answer existed that we could all agree on, it would have been found by now. The ideal goal is for us to have enough good candidates running and succeeding. How do we make that happen? By keeping an eye out for good candidates and by !voting responsibly. Of course, one positive aspect of the current situation is that it is very hard for a poor candidate to pass at RfA. We don't want to over–correct and start letting those poor candidates succeed, but the present status is hindering quality candidates, and thereby hindering Wikipedia. The candidates don't want to run because of the negativity surrounding the process, and I would argue that some of the discussion that attempted to improve RfA but went sour has made the process even less appealing.
All we can do is support the good candidates that do run and encourage more to try. If we ask enough of them, eventually somebody has to say "Yes", right? RfA has a public relations problem among prospective candidates and the best way to fix this is for good candidates to get good results. The problematic lack of candidates will not be changed by unfruitful discussion but rather by hard work.