As for the third link, search ads are not the same thing as mobile display ads from the user's perspective. Remember that in the other case (mobile display ads), Google is probably selling those ads, too.
There are other reports out there & some of them do great breakdowns of click volume, click costs, click values & so on ... based on things like: device type (desktop vs tablet vs cell phones, or even across operating systems), branded keywords vs unbranded keywords, text ads vs PLAs/shopping ads, etc.
There are many such reports, but the SEM company which regularly puts out the best report on this front is RKG (which was bought out by Merkle last year http://www.merkleinc.com/news-and-events/press-releases/2014... and even after the buyout is still putting out great reports each quarter).
RKG's most recent report is downloadable here http://www.rimmkaufman.com/digital-marketing-report/ & it is something I read (and recommend) every quarter. :)
In the report, they stated they saw Google dramatically increase minimum CPC on own brand keywords over the past year, that brand keyword cost was up 39% YoY, and that Google was now charging brands more for clicks on their own brand terms than Bing does. You can see a quick review of their report here http://searchengineland.com/us-paid-search-growing-but-at-sl...
And there might be multiple factors in Google showing fewer ads. Some general assertions and/or hypothesis on potential drivers...
A rise in minimum CPCs on brand keywords might block out some of the lower value 3rd party arbitrage on the branded terms, Google might be showing larger ads (with more ad extensions) above the organic results & fewer ads in the right sidebar, some categories might have PLAs / shopping ads performing so strong that there is less need for text ads, in some categories (like hotel search) Google has effectively turned much of the "organic" search result set into another layer of ads via their hotel price ads, Google includes affiliate links in their knowledge graph, over time more of Google's search volume is on mobile devices with a smaller SERP interface & no right rail to show ads in, over time more of Google's search volume is in emerging markets with less shallower and less mature ad markets, etc etc etc
A few other things worth mentioning on their general economic trends: on smaller devices partners (outside of the big one in Apple) are easier for Google to squeeze out & represent a smaller share of Google's ad revenues, over the last couple years the partner network keeps comprising a smaller percent of Google's overall ad revenue (numerous sources on this front, like http://www.siliconvalleywatcher.com/mt/archives/2013/10/anal... & http://blog.pagefair.com/2014/adsense-smoking-gun/ ), Google is getting higher revenue from clicks on Google.com & their regional search properties but the YouTube pre-roll ads generally go for less money (since they are earlier in the consumer funnel than search is & have less intent than a search does) & drag down the aggregate click price (because Google counts YouTube video ad views as clicks).
One other thing worth mentioning in terms of mobile is that while the mobile click values (and thus CPCs) are lower, Google offsets that by dominating the mobile interface with ads. 3 years back I wrote this post http://www.seobook.com/mobile-seo highlighting how Google had over double the CTR on mobile search ads as they did on desktop search ads. The top graph from that post came from an RKG quarterly report for that quarter.
https://twitter.com/danbarker/status/439125570115223552/phot...
And they link directly to Wikipedia, driving traffic to their site, so they're not really wrongfully stealing the data.
However, for sites that do monetize through ads, Google's 'scraping' of the content is indeed stealing if the content would 'organically' draw more clicks otherwise.
It makes sense Wikimedia didn't see an immediate donation drop. In terms of a conversion funnel, the people who are using Wikipedia for the first time are not likely to donate. My guess is most of the people who are most likely to donate a small amount here or there are the regular users who regularly seek out the site. The drop in donations due to the lower exposure would take years to kick in, because now with 20% or 30% fewer users, there will also be less new future power users who would donate in say 2018 or 2019. But offsetting any decline there for a company making over a billion a month in profits shouldn't be too substantial. Sergey Brin donates a half-million a year grant to the Wikimedia foundation & Google has also donated a few million.
It is also worth mentioning the knowledge graph now often contains affiliate links for things like music, ebooks, and so on. Over time more and more knowledge categories will have paid affiliate links in them (say booking a restaurant, heading to a concert, etc.)
In some cases Google may temporarily make a link type free to try to drive adoption and awareness, but over time it will be the same sort of scrape-n-displace they've done with hotel bookings. They not only turned the "organic" hotel search listings into a second set of paid ads, but now they are also testing having some users convert and buy the hotel booking while on Google.com - just like they are doing with their car insurance offering. And they have also added links to cityname hotel searches in their knowledge graph in order to drive more people into that "almost nothing but ads" funnel.
You can see the net effect of that sort of displacement on an industry by looking at all the consolidation in travel. Expedia acquired Travelocity and Orbitz. That leaves Expedia, Priceline & TripAdvisor as the remaining 3 big players. And even with TripAdvisor Google offered them and Yelp an ultimatum to allow Google to scrape/steal their reviews & if they wanted to opt out of that theft then they could use a robots directive to block Googlebot from indexing their site. That monopolistic abuse was so brazen it actually managed to draw the interest of regulators and politicians. Eric Schmidt did a fantastic job of misdirection & fibbing at the 2011 congressional hearing (ref http://www.benedelman.org/news/040115-1.html ) but the regulatory & political reviews forced Google to back away from at least that form of monopolistic bundling.
Though, of course, the same sort of behavior continues to this day in other forms. From just yesterday there's both https://twitter.com/chrispirillo/status/624630005734113280 & https://twitter.com/jeremys/status/624675646694780928
* http://www.zdnet.com/article/googles-mobilegeddon-moves-hitt...
Perhaps Google is trying to boost CTR / relevance in order to attract/retain the more valuable advertisers (who would probably otherwise weight Facebook more highly for its better targeting). I really wonder how much of this growth is just Google (and Facebook) extracting more marketing dollars from the biggest spenders.
Edit: looking over this thread, there are several rather mild comments that got strong downvotes. Looks like the followers of this thread were indeed inordinately pro-Google.
So, iow, business as usual.
It's no surprise that after the "mobilepocalypse" earlier this year we're seeing mobile profitability grow for Google.
Google effectively forced advertisers hands in bundling their inventory and is agressively raising min CPC in a number of categories (2)
And they are breaking many of their OWN rules about how aggressively one can promote advertising on mobile - excellent, and entertaining breakdown of that here -> http://www.seobook.com/google-goes-mobile-unfriendly
I don't have much of a point other than sharing some of the information behind what's actually driving these earnings, and the major changes I've seen / experienced as a guy that spends a lot with the GOOG.
(1) http://www.wsj.com/articles/googles-results-top-expectations...
(2) http://searchengineland.com/google-showing-fewer-ads-per-sea...
(3) http://www.seobook.com/google-goes-mobile-unfriendly