Can you please let me know whether I am wrong. It seems to me that the first argument only makes sens if there is a limit set by the person who matches, and if there is no choice in charity.
I am asking this because my current employer does gift matching, and I was trying to find out what EA says about this kind of action. Your post is quite interesting, but I am far from sure it actually apply to the matching-donation they do, and I'd like to see whether it is because your argument only works in some context which were not explicited here.
However, I have an important choice of charity to which I can give, trying to solve very different problems. I would agree that thinking "I ought to give to this opera house because my gift will be matched" would be a poor choice; after all the impact I could have by giving 2N to an opera house is far smaller than the impact I'd have by giving N to an EA charity. But if I can actually give to funds helping people, even if they are not fundations evaluated by EA associations, it seems to me that giving them 2N $ may be better than giving N $ to a classical EA charity (assuming that they are not more than two time less efficient than the EA charity)
Actually, even if the employer already know that they want to give N$ to charity this year, it seems to me that influencing their choice and ensuring they give to more important charity is worth considering (I would not be surprised if it was actually a reason why they have this politic, to ensure that their gift is aligned with the one of their employees)
Furthermore, as my employer has not put a company-wide limit on the matching, it seems that It would not be dishonest, if all employees given the maximal matched amount of 10 000$/year, by its policy the company would actually give hundreds of millions$/year. So it seems that in this case, matching-donation is less dishonest than in the case where there is a limit and the donor already have a fixed amount they want to give in their head.