Drug Monkey has a great post on from whence does the over-glut of people looking for an academic job/grant come? It is an answer to a post from Thoughts for Breakfast about the morality of hiring postdocs. Needless to say, I come down on DM's side of this argument, which in a nutshell is:
BUT THEY [postdocs] DIDN'T GET DROPPED OFF BY THE STORK OF SCIENCE!!!!!!!!!!!!!!!!!!!!
One of the comments, from karassment, deserves more consideration (slightly edited for space):
I have been looking for a TT job for years, and now that it finally looks like it will happen, I think about this all the time. I am acutely, painfully aware of how hard the job market is. My current PI has flatly refused to take any students, .... I am not going to have the clout to do that, and even if I did, my start up will disappear really fast if I hire only high paid uber techs and staff scientists like you said. The students I'll be talking to about joining the lab will already be students, [more stuff on working with PhD students]. I'm not going to have much access to undergrads, and don't think I'll have the budget or fame to attract postdocs for a long time [my bold]...
Karassment's comments, the ones I redacted, about PhD students are heartfelt and admirable.
What I want to address is the point I bolded: places to look for lab-peeps without feeling you are being dishonorable and immoral for training people who can't ultimately get the jobs for which you are training them.
In my various past (often evil) lives, I did not have good access or opportunity to get grad students. There were no PhD programs in two of the departments to which I belong/belonged. Another dept had a tiny PhD program, where stealing of students from other labs was time honored, jolly old boys BSD tradition (which when this was brought up at a faculty meeting, the condescension would have done Lady Catherine de Bourgh proud). By the time I was senior enough to attract grad students, I saw the problem and didn't think it right to take any.
So what can be done in this situation?
First, do not overlook undergraduates wanting to do research. Many are pre-med, or other pre-clinical programs. I have found that these folks can be hard working. While not PhD students, they can be productive in ways that jr faculty need. They will undoubtedly need more supervision, but that doesn't mean that can't be productive. They can generate data, and write up a report that can be the basis of a paper. What is important is that they have a project that they can own, and that you talk to them minimally 2 times per week to keep them on track. Biggest mistake with such students is giving them too much rope, as they will always hang themselves. And, even if you don't have access to UG's at your uni (as Karrasment suggested), its worthwhile reaching out to another nearby institution to see if you can forge a working relationship.
Do not necessarily assume you can't get a good postdoc. Keep your ears & eyes open. Maybe not the first year, but the second can work. People chose grad schools on The Name of The Uni basis, and sometimes on the program. People chose postdocs on the basis of the mentor. Attending meetings in one's subdiscipline, one can approach current grad students and say "your work is interesting, let me know if you're looking for a postdoc", and give them the 2-minute schpeil on your work. Keep in mind, at some private MRU's a student costs as much as a postdoc.
Another source are students in master's level or PhD clinical programs - SLP's, PT's, AuD's Dental students, Dental residents, some nursing programs, psych - folks that have do some kind of thesis. Often these schools don't have enough research faculty to supervise all the students they need. A friend of mine was responsible for 10 graduating MSc students a year, which in a 2-year program was hell. She was tremendously grateful that I took two of these into my lab. Again, they need tightly supervised projects. Part of the skill in working with these people is to craft contained projects that are useful to your greater vision/goals, but are interesting to them. You may not have these schools at your uni, but there might be some nearby. It's worth looking into.
Again one of the keys to this is to understand how these students are different from you (the former PhD student). They are not living and breathing and excreting science they way you did. They are doing something that they perceive needs to be done to finish their degree. But, they can also perceive that working in your lab maybe doing it in a more interesting way than the alternatives. You don't have to entertain them, but you do have to treat them like human beings. More work on organization and supervising is critical to getting what you need from the interaction.
Finally, if one of these doesn't work, end it quickly. A student not showing up, doing something bad with the animals, screwing around with the data, etc, is not worth saving. Learning to move on is also important.
I don't think that Karassment's situation is dire. You don't have to give into the soul-sucking strategies for success. I also think that having one grad student at a time, someone with whom you work to ensure a higher probability of ultimate success is defensible for younger scientists.