The figure of 103,400 deserters cited is, I'm sure, based on documentation that appears to be authoritative; that is, until the documentation is looked at with closer scrutiny. When I first began transcribing the service records of Arkansas Confederate soldiers, I dutifully recorded the references to desertion. However, when I began to put together a database, wherein the service of a soldier in multiple regiments were consolidated for the first time, I noticed that the majority of those soldiers reported as deserters in their regiments, were in fact serving in a subsequent regiment. I guess you could say that one regiment's deserter is another regiment's volunteer. Because of the way the Compiled Service Records are organized, it would be a monumental task to arrive at this revelation. It's only when you put them all together in a database that the true picture emerges.
By this, I don't mean to minimize the detrimental effect that desertions had on the Provisional Army of the Confederate States. And it only got worse as the war ground on. It is amazing to me that a "rebel" army, organized from scratch, opposing a much larger, stronger and better equipped army, managed to hold together and fight so tenaciously for four years, without triple the number of desertions it reported.
However, my research, stretching over more than 30 years, has made it clear that the number of "real" deserters, i.e., those men who left the Confederate service for good, is somewhere between a third to less than half of the number of all reported deserters. Technically, men reported as deserters were in fact deserters; that is, they weren't where they were supposed to be at the time they were supposed to be there. But if you look at the broader picture, most of them were somewhere in Confederate service. That's why I always caution people not to end their research when they come across a report that their man deserted. He may have deserted that outfit, but the odds are that he joined up somewhere else.
One caveat -- the vast majority of my research has been centered on Arkansas soldiers, including those who served in regiments from other states, so I can't say with any degree of authority that the same results will be found in other Southern states, but I strongly suspect that things were pretty much the same, with the possible exception of the border states, for entirely different reasons.