A link theory that gets thrown around a lot is the concept of SERP juice. Is juice the best analogy for how linking and voting really works?
Evidence for and against Link Juice
The PR system. Certain links are worth more than others. There is possibly a sub-PR ranking that determines how much juice you're getting sent.
The effectiveness of Link Wheels. If Links operated like juice, then two links to your "money" page would be the same as a link to the link of your money page. In fact, two links to your money page would be better because you would get the anchor text benefit. But, the thing is that if your links aren't indexed then they won't count. However, link wheels usually include authority sites like Ezine Articles that already get automatically indexed.
The Power of Content. All link juice stems somewhere from content. Even with pages without a lot of content their link juice was developed from content. There's tons of anecdotal evidence that content surrounding links increases the amount of link juice that page gives out. Ineffective links aren't within original, unique content whereas effective links are. Content produces link juice which floods out in your links.
The diminishing of links from the same C-class IP. You can minimize this effect somewhat with on page optimization(i.e. write more content) and by varying this anchor text. And the effect can't be that drastic or links to sites like ezine articles would be ineffective. Their is a way to account for this in link juice theory. We already know that no-follow dissipates link juice. Everything just places a modifier on the link juice that goes through and the rest dissipates. Things like relevancy, C-class ip, etc. make the valve smaller for the link juice to go through. Things like trust ranking and authority status add more link juice to go in. So some things affect the link juice valve whereas others affect the amount of link juice.
Google Employees have used the term Link Juice although it is a lot catchier term than Link Votes. The coolness of the name may be an explanation for it's usage rather than it being a valid description of how links are used to rank the site.
Link Juice model encourages Black Hole SEO(where every link stays within your properties). If link juice flows through the valve system and sites are ranked by the color of the juice(the color of the juice being affected by content and the links pointing to the site) then it makes sense to not link to anybody else unless you get an advantage. But, that is not the case as linking increases relevancy factors. It is a lot like bolding your keywords where you get extra credit for the anchor text you use to link out to other sites. If the link juice model were correct than sites like Ezine Articles would not be a PR6 as they would have massive amounts of link juice flowing out of it. Ezine Articles has lots of internal links on each page as well but with the amount of Ezine Articles written per day versus the amount of internal links Ezine Articles has per page than link juice would dissipate from Ezine Articles at a rapid rate!
Conclusion: It's not link juice but content juice. If their are only two sites on the internet with no links to either of them and everything else is equal how do you rank them? By content. Content is the primary determinant for how much link juice a page gives with modifiers like relevancy, etc. One contra-indication to this is the lack of duplicate content penalty. If content determines how much juice a link has then you don't want people copying a whole bunch of content just to get more content juice. It's easy to spot a spam blog but it's hard to spot a duplicate content blog.
So, how do we determine how much "link juice"( I can't think of any other name for it!) a link really gives out? Google Webmaster tools tells us a lot. It tells us unique anchor text is important as it lists every single anchor text your site gets. It tells us that how much the sites linking to you gets crawled are important as it doesn't list links that don't get crawled. It tells us that deep links count extra as it shows the links to each page individually. It tells us that internal links matter but it only gives internal links numerical votes as it doesn't display anchor text. If Google Webmaster tools is believed what matters is Anchor Text, quantity of links, quantity of pages linked to, number of internal links, plus keyword score. Of course, the people doing Google Webmaster tools could be totally independent of those determining the search engine algorithms so it's all just speculation.
So really what's the best analogy for link juice or whatever. Google is like your parents. Their house, their rules. How much the link counts is however much Google says the link counts for.