Writing from personal experience, it seems that regardless of which CAD systems I can use, the contract position I am applying for always seems to be specifying skills in a different system.  A variation of Murphy's Law always seems to kick in.  Although I stress that I will take a crash course in the new system, the employer wants (and is able to get) a contractor with the skills to “hit the ground running”.
I expect that eventually the pendulum will swing in my favour, but at the moment it is well over the other way.  Maybe I'm just being paranoid!

I started my working life in the drawing office and I remember a sea of drawing boards with machines of various vintages.  The apprentices always got the oldest, most knocked about machines and were always ribbed and criticised for not drawing square lines.  Those old counterbalanced parallelogram machines were pretty difficult to hold square,  but it was a good learning environment.  
In those days the majority of drawing office people smoked and I can remember a constant fog of cigarette smoke.  Air quality was an unknown factor and no one would ever dream of opening a window.  I'm not even sure that the windows did open.
The drawing office was divided into areas for engineers, designers, detailers etc.  Everyone had a drawing board, from the senior design Engineers down to the newest apprentices.  

As the level of work ebbed and flowed it was relatively straight forward to bring in contractors and for them to “hit the ground running”.  
Having drafting skills meant that contractors could readily move between the different disciplines.  Mechanical, civil, structural, electrical, the fundamental skills required to produce drawings were generally very similar.   Perspectives, line weights, dimensions.  Whilst there were some differences in terminology and styles,  the fundamentals were very similar and contractors could generally move between the disciplines fairly readily.
Going back about 20 years,  Autocad dominated the desktop market and it was a similar situation  with most contractors having Autocad skills.

I got this far into the article and then I started to go off at a number of tangents and soon realised that there was the potential for this to be bigger than Ben Hur, if I went into detail with all of the points that I wanted to highlight and explore.

For example there are the higher end packages, Pro Engineer, Catia etc, then there are the mid range products like AutoCad, Microstation etc, then there are the low cost products, then there are the “Free” products, and I use the word “Free” in its's broad capacity as many of the free offerings have caveats of some sort.  
So where am I now, oh yes, lets keep going.
Was that 2D or 3D, Ancillary Programmes which are industry specific, Rendering, Animation, Calculations, Presentations, Compatibility, Windows, Mac, Linux and by the way, “do you want fries with that”.

Also, although maybe not directly related to the initial thrust of the article, it would be worthwhile to consider how companies select a CAD system, as this can impact decisions by contractors to learn new CAD systems.
There is the option to look at market share and “just go with the biggest”, or to be industry specific “thats what everyone else in this industry uses”, or functionality “the 3D capability is excellent, but will the specific industry only ever need basic 2D”, or "3D plus render/graphics/animation are absolutely essential", etc, etc, etc
Lets not to forget that a CAD system needs to run on a computer/monitor package with the correct specifications together with a large format plotter. Remember that the IT department needs to “understand” CAD as the requirements are very different to the simple word processing systems. A computer which is under specification will cause continuing frustrations and costs due to factors such as slow speed or crashes.
It would also be good to touch on the fact that the initial investment in CAD is only the first part of a long term commitment (read “expense”) as there will be a need to keep upgrading when each new version of the software is released, as inevitably, each new version always needs hardware with higher specifications.  
Of course each new version also introduces increased levels of specialisation and knowledge which means that the CAD operators are more specialised and it is then potentially more challenging to quickly and easily find short term contractors.
For long term projects it may be viable to train contractors, but for many applications the need is for short term labour able to “hit the ground running”.
Oh yes, this is the point I was making at the start of the article

I almost lost my train of thought for a moment, lets get back to the CAD software.  
Again, maybe not directly relevant but worth noting is that there are also different types of purchase and licensing arrangements.  In some cases the user doesn't actually own the software, the user is purchasing the right to use it, but it is never legally the property of the user, to (say) resell. Or the CAD system could be rented with a periodic fee which includes upgrades. There are numerous options.

I could also delve into the ethical side of CAD.  Pirate versions can generally be sourced and whilst I would never advocate using a pirate version commercially as it is strictly illegal, I can appreciate a reluctance by a contractor to pay several thousand dollars for a legal version of a specific CAD system in order to learn at home.  Maybe a philosophical company would turn a blind eye to that situation, knowing that having learned the software at home, the Engineer/Drafter would then be applying their newly learned skills in a commercial capacity in a legal state.  
Whilst there are many companies and educational institutions teaching CAD, there is no substitute for hours in front of a computer in actual design and drafting work. Whilst a training course is useful in order to be aware of a CAD systems capabilities, it generally will not equip the user with the skills to be proficient to the level required by an employer.

Summary
So where are we now.  Well, the irony is that CAD brings such enormous benefit to Engineering in so many ways, but the large number of competing CAD systems has the  potential to create much less flexibility in the labour market for both Contractors and Employers.  
Having said that, many of the CAD suppliers are very small and by applying the Pareto Principle it should be possible to narrow down the number of options. I guess that the experienced contractors know which systems are more popular in their particular fields.
In closing, all I can say is that in my particular case, Murphy's Law beats The Pareto Priciple every time