It depends on the music you make. If you're combining various instruments at once, then a mixture of stereo and mono settings would be the most effective. But if you're using only one or two synthesizers, then I would highly recommend stereo.
Considering the often immense sounds we design on polyphonic synthesizers, it is quite unnatural to then squeeze their overall sound into a little box called a speaker using a mono signal. It's like listening to a symphony on a pocket radio. Sounds come with their own acoustic environments, which include various effects, the most important one being space. This is true even for melody instruments. Stereo is a minimalistic imitation of this all-important effect, but it's certainly better than no spatial effect at all. With larger types of sounds - strings, brass, organ, choir, or ethereal pads - it just makes musical sense to use the orchestra and pipe organ as models for emulation. These ensembles/instruments occupy a large amount of physical space - a stage or a wall - and as a result, specific tones are heard at different locations across this space. This contributes to their enviable depth and richness. Again, it just makes good sense, in designing sounds that are reminiscent of such ensembles/instruments, to strive for a similar depth and richness by somewhat imitating the spatial effect through stereo.
In my own synthesizer patches, I maintain a constant stereo field, with very few exceptions. I consider a string, brass, organ, choir, or ethereal pad in mono to sound terribly artificial and electronic.