IF released by Thomson Reuters (JCR) is always considered and accepted by all as an authentic IF. There are many nowadays who releases their own impact factor like Universal Impact Factor, Scimago etc.
I would recommend reading this [ http://scholarlyoa.com/2014/05/20/impact-factor-confusion-spam-emails-mislead-researchers ]. There is only one IF and that is the one from Thomson Reuters. I would suggest reading the information on their website [ http://thomsonreuters.com/journal-citation-reports ].
There are an increasing number of publications discussing pros and cons of impact factor (IF) which I do not want to summarize or comment here. But one important aspect of your question (who is the deciding authority?) concerns to the scientific forum in which we are moving (ResearchGate, RG). Can we trust the information on the impact factor (of a certain journal article) and on the impact points (in the researchers profiles)? Who and above all, how gets RG these ratings? The IF is specific for particular volumes and for a certain period (2-year-impact vs. 5-year-impact) and it changes from year to year. This means that journal articles from different volumes of the same journal/ periodical will have different IF. But at RG all publications from one journal have the same IF. Thus, the accuracy of the IF for most of the researchers at RG must be questioned.
Dear Sultan Salah, this is an important topic in scientific publishing. Therefore I would recommend you (or request you, respectively), to add "publication ethics", "scientific publishing" and "academic writing" to the topics of this question.
Dear Blake Gurfein, what we really need, would be a page which displays the older IF, too. Only very few journals or publishers publish the IF of several volumes.