Thats a really interesting question actually. I have found that you can never guarantee what you are going to rank for.
At issue here are the URL, titles, anchor text on the links and keyword density in the content. I will call these keyword elements for convenience.
I have created several wheels where I used only the exact keyword for all keyword elements. Some ranked on the keyword in G, some ranked on other keywords but not the target, and some have yet to rank at-all. What's interesting is that I got the same result when I varied my keyword elements in different wheels - for example, using squidoo.com/*****news and hubpages.com/total*****
For one exact-keyword wheel, I failed to rank at all for the exact keyword but ranked on similar long tails that still brought in great targetted traffic (see thread below)
For another wheel, I was targetting "xxxxxxxxx promo code" and mixed around my keyword elements quite a bit. Not only did I succeed in ranking for my target and countless long tails, I also ended up ranking for the manufacturer's own keyword by accident even though I had not targetted it, only its promo code deals. This has happened a few times and brings in big income. However, all the times I have pointed my keyword elements toward a 1 word manufacturer I have always been unable to dominate - so far.
So, the evidence seems to be pointing very slightly in favour of using some variation in your keyword elements. But, thats not to say that you won't succeed with an exact-keyword wheel in the future, and I am going to keep trying them every now and then.
This thread should be of some help to you:
http://www.blackhatworld.com/blackhat-seo/black-hat-seo/100159-unusual-linkwheel-effect.html