Great for refereeing, but the bulk of our service work is committee work for University, College, and Department. So less useful as a rubric for “service” all told.

Maybe “Norm for how much refereeing…?”

This sounds to me like a good rough rule of thumb. But I also think it makes sense for different people to play to their different comparative advantages. Some people should do more refereeing than this if it’s a thing they don’t mind as much as other people do and they are pretty good at it, and they might do less conference organizing and department chairing, while other people should do more conference organizing and less referring or whatever.

This is roughly what I already do. Probably not the “plus 1” part, but that’s only because I don’t get asked to review very often. I accept almost all review requests as it is for the same reciprocity reasons as in the post.

On the refereeing thing, if we followed the OP’s suggestion of reciprocity plus one I suspect that the refereeing system grind to a halt. Here’s my critique of a similar suggestion made by Caterina Dutihl Novaes on New Apps in 2014. What Catarina was suggesting was a similar principle: that if you are a fairly established member of the profession (e.g. not a graduate student, adjunct etc.), and you submit X articles in a year, you should do 3X referee reports: 2X to cover for your own submissions (each submission generating two referees’ reports) , and an extra X for the surplus of people not in a position to act as a referee themselves. ‘Does that sound reasonable?’ she asks. Well in one way it does and in another it doesn’t. It would be a reasonable refereeing workload for an established member of the profession and if that were all that they were expected to do they would probably do it promptly. The problem is that things are currently arranged the system would collapse if academically successful philosophers did not do a great deal more than that. And it is *because* they often get asked to do a lot more than that, that refereeing burn-out sets in, that the as-yet unrefereed papers pile up and that the reports when they appear are often snarky, superficial and dismissive.
First point to note. The top fifty journals have rejection rates varying between 80 and 95%. This means that the vast majority of PUBLISHED papers are rejected the first, the second and even the third time around. This suggests that the average PUBLISHED paper has generated at least six referee’s reports and that there are probably quite a lot of papers that never see the light of print but which have generated many more.
Second point to note. You get asked to referee a paper under one or more of three related conditions: 

1) you have published something in the area that has garnered a couple of citations or has otherwise created a stir,
2) your own work is cited in the paper to be refereed,
and
3) you have an established reputation as a reliable referee.

You don’t have to have to have published all that much in a given area to meet conditions 1) & 2). For example, I have only one publication on truthmaker theory (a coauthored paper with my colleague Colin Cheyne), yet I am often asked to referee papers on truthmakers and negative facts. Obviously you don’t get to meet condition 3) unless you have already done a bit of refereeing, usually, though not necessarily, for the journal that makes the request.
Now imagine a successful philosopher, call her Sophie, all of whose papers get accepted the first or the second time round. Thus the average number of referee’s reports per published paper that SHE generates is three. Since speedy acceptance is at least roughly correlated with quality, her papers are generally good and are regularly cited in the area or areas in which she works. This means that she often meets conditions 1) & 2) above. Let’s imagine too that Sophie is a conscientious person eager, at least initially, to do her bit for the profession. So she soon acquires a reputation as a reliable referee. Pretty soon the referee requests come flooding in. She gets a lot more than the nine requests per published paper of her own that Caterina’s formula would suggest.
Now imagine a less successful philosopher, call him Jack. None of Jack’s papers are accepted the first or the second time around. The average number of referee’s reports per published paper that HE generates is nine. Furthermore, since tardy acceptance is at least roughly correlated with poor quality, his papers are pretty ho-hum and are NEVER cited in the area or areas in which he works. (Remember that four years after publication, the median number of citations per philosophy paper is ZERO – that is most papers are not cited AT ALL.) This means that Jack NEVER meets conditions 1) & 2) above. Because he never meets conditions 1) & 2) , he is unlikely to acquire a reputation as reliable referee even if he is a conscientious person with a knack for seeing what is good or bad in other people’s papers despite deficiencies of his own work. Is Jack really going to get the 27 refereeing requests per published paper of his own that he would need to complete in order to meet Caterina’s formula? Obviously not.
Thus the problem I suggest is this. For structural reasons, the burden of refereeing falls disproportionately on the more successful members of the profession who are the very ones who generate the lowest numbers of referee’s reports. If they confined themselves to reciprocity plus one, then the system would fall over very quickly. There just wouldn’t be enough willing and qualified referees (or willing and identifiable-as-qualified referees) to do the necessary work .