Most of the time, I record a song completely and then bring in a singer to do the vocal(s).  These projects often have orchestration and a fairly large track count.  Recording the vocal has two problems: 1-Latency and 2-adjusting the mix on the fly so the singer hears what she needs.  Here's a solution that's been working for me:
 

1. Make a copy of the project (Example: Save As "Song 1 Vocal Record")
2. Mix the tracks to a single stereo track and then delete all the other tracks and virtual instruments, and add a few blank vocal tracks.  (I usually create one vocal track with the right input and routing and then clone it a few times.)  Do not add or remove any bars or make any tempo changes.
3. The result is a single stereo mix of the music with some blank vocal tracks and no virtual instruments or unneeded plug-ins.  At this point, I usually lower the ASIO buffer to reduce latency further. 
4. I do all the vocal comping, Melodyne correction, etc. in this project.  When I am happy with the vocal, I bounce it down to a single vocal track and copy that into the master project.  (I save this project, of course, so I can always go back.)
5. If you do this, be sure to copy the entire vocal track from the beginning of the project or else be sure you have a clear marker at the start of the vocal clip (e.g., Bar 6, beat 1) so you don't have to worry about nudging it into place. [On a related note, I always trim main vocal and harmony clips so they all start at end at the same place and on a specific tick (e.g., beat four and a half before the chorus).  This makes moving, doubling, etc. easy.]

Of course, this requires a little prep work, but it makes the vocal recording session a snap.  Besides eliminating latency, I don't have to fumble around trying to turn down the horn or bring up the strings or get the singer's voice loud enough, etc.  I only have to manage two faders.

yup.  i often follow this method.  to the tee

I can see how this would help if you're monitoring through your audio converter, however, if you monitored through a mixer then latency is not a concern when recording vocals unless you're trying to hear f/x in real time.  I don't care about hearing f/x when recording a vocal, just at the mixdown stage.  This seems like a lot of unnecessary hoops to jump through, but if it works, then OK.

I'm with Lynn here, this is why it is useful to have some form of reverb earlier in the chain, whether it's a hardware one or an audio interface with built in fx.

Still, if you don't have that option, I understand the appeal of this workflow!

All you need to do is do the mix down to a stereo track in your project and then toggle the FX.  That should remove any latency issues as I do things that way but at times I'll mix a couple stems to have a little more control in case the singer is picky about what they want to hear.   Just solo the mixed tracks you want to hear.  Overall, Vocalists like to sing against a more exciting sounding mix when possible so I'm with you regarding providing a rough mix (or whatever stage you're at in mixing the track) so I'm on board with the method. 

An alternative ...
Bounce the mix. Route the bounce to a direct out, not the master bus.
Goto console view and add pro channel power buttons to a group and the mute buttons to a group. Don't include the bounce.
Goto track view.
Click a mute button on a track to mute all.
Click a power button on a pro channel to disable all.

Reduce sample buffers.

Record vocals.

I don't see why you would have latency if your system is set up properly for direct monitoring.
And if the singer really needs reverb to perform at their best then you need the proper front end hardware to accomplish this.
There is a way I found a long time ago to add reverb using the input echo.
Still monitor direct at the interface. Now turn the effect send on the track up full but turn the tracks volume way down until you don't hear so much delay. Kind of a hack but it does work better than hearing the delay.

I certainly understand this workflow very well .Although I'm not a singer nor do I work much with singers . I use this type of workflow for other creative reasons .
I pretty much do the same exact thing when it comes to laying down all my lead guitar tracks and the occasional midi guitar synth parts .
I always like to have a parent project & a child project .
The child project  is wide open containing a mix of everything I plan on keeping for my song project / track .
Working this way the child project always seems much snappier for a few reasons .
All the plugs and sounds I plan on using for this premix / work-file were already rendered and my GUI is wide open .
Having a wide open GUI fits and works for my emotional temperament ...
Having to contend with clutter and distractions on my computer screen seems like a  good excuse for the Muse to go away and take a long vacation and never come back 
Since I don't want that to happen I bow down to her Mighty Holiness and stand there with a large palm leaf to fan her gently as I peel her lunch time Grapes if have to 
Baby please don't go , Baby please don't go, Baby please don't go down to New Orleans because I love you so
 
For SONAR to SONAR projects and playing transfers this works pretty good ..I may even decide to open up the Parent project and copy over my newly created performances over from the child project in one felled swoop. 
Ultimately everything including The Good The Bad and The Ugly will end up in my SONAR Parent project anyway .
 
I certainly want it all there
FWIW , I don't seem to mind all the pre work and planing that may be involved . I'm OK with that .
A small minor RUB can be keeping a very close eye on file management and keeping a tidy musical household ...
With SONAR to SONAR this is not such a big deal for me ...
 
Sometimes I use other brand DAW's that have all their instruments , plugs and other doodads locked strictly to them ...
For example , Logic has some real nice sounding synths and plugs , same with Samplitude and Abelton Live ..
When I want to accesses some of the synths or locked only to the brand of DAW sounds .
I don't consider it such a big deal to send my mix from my parent project over into some of these other DAW's to be able to work with and to be able to access sounds I cant access in SONAR only...
Now some people may feel this is a major waste of time ...personally I don't ....
The day I can afford the money to go out and buy every 3 rd party sound sample , synth library and plug that I can stuff into SONAR that I feel  will make me happy , That will be the day I will be able to keep all my work strictly SONAR ..
All these other DAW's export audio and I always bring my work back into SONAR anyway ...
Once again the only major RUB I have found is keeping my file management issues in check and keeping a clean concise track of where all my files and projects including the children are ...that in it self can be a lot of work ...
anyway that's my story and I'm sticking to it
all the best,
 
Kenny 
 

I do the mix down to stereo and applying a couple effects for the singer has always worked well, not enough latency to tell.

stxx
All you need to do is do the mix down to a stereo track in your project and then toggle the FX.  

Absolutely. 
 
OP seems to be trying to invent the square wheel here.