John Robinson devotes Chapter V of Redating the New Testament to the Epistle of James. Dating it to c. 47-48, he argues that the epistle is likely the earliest Christian text still extant. As we saw in his discussion of the Synoptic gospels, this gets a little tricky, as he holds that a "proto-Matthew" and a "proto-Mark" likely preexist c. 47-48, but puts our gospels of Matthew and Mark proper at c. 60. These proto-text hypotheses become important for his argument, as he argues that the Epistle of James reflects something closer to the early stages of the Synoptic tradition, as he understands it, than to the later stages. I'm less-than-sanguine about such hypothetical proto-texts, but nonetheless I am persuaded that Robinson is on more or less the right track with regards to the Epistle of James.
We can see the strength in Robinson's position if we compare it with that advanced by Dale Allison in his 2013 International Critical Commentary on James. Now, I need to be clear: I think Dale Allison is one of the best NT scholars working today. His careful attention to detail and intellectual integrity make him second to few, if not none. Yet on the question of James's date, I think that he is mistaken. Like Robinson--and indeed most commentators on the epistle--before him, Allison recognizes that the text reflects a Christianity very much marked by the movement's Jewish origins. Indeed, as Robinson emphasizes, there is not even a hint that there are Gentiles involved in the movement at all. The letter, although clearly Christian, envisions a Christianity that is wholly Jewish. On this, Allison and Robinson are largely in agreement, but they differ greatly in what they infer from this agreement. Robinson infers that this points towards a time when Christianity could be conceived as a wholly Jewish movement. Since it is difficult to envision any Christian anywhere thinking in such terms much later than the council of 48, Robinson points to a date in the late 40s. By comparison, Allison infers from the Jewish-Christian character of the Epistle that it was produced by a second-century Jewish-Christian group, such as the Ebionites.
Here I think that Robinson has the stronger argument. Allison's argument for pushing James so late relies largely upon the lack of clear external attestation for the letter prior to c. 200. Surely, Allison reasons, if it had been written earlier it would be attested earlier. I'm not altogether persuaded. Given the fragmentary nature of our textual witnesses from this period, such lack might not in fact be that significant. I'm not convinced that attestation is particularly probative data for establishing the date of our epistle, apart from setting an absolute terminus ante quem c. 200. Perhaps more crucially, I have difficulty envisioning a scenario wherein Allison's Ebionite James ever makes it into the NT canon. This of course is what sets it apart from other known texts that might have issued from similar circles: it is a historical datum that the Epistle of James ended up in the canon, and thus one must account for that datum. Allison has to envision a form of Christianity so distantly removed from those that apparently generated the canon that it can altogether ignore the presence of Gentile believers in their midst, and I struggle to see how a pseudonymous text from such a form of Christianity made it into the canon. Precisely to the extent that Allison must emphasize the theological and social distance between those who produced the epistle and those who received it into the canon, to that extent he vitiates our capacity to account for that reception at all.
By comparison, Robinson has a ready-made explanation for the canonization of the epistle: it was preserved because it was remembered to have been written by James. If one wants to account for the absence of attestation before c. 200, which might or might not be probative anyways, one can look at the very fact that it does not address a situation that seems to have attained after c. 50: although James's prominence ensured that the letter was preserved, the archaic content rendered it of limited immediate relevance to the life of the developing church. The only other option I can conceive to preserve a later first or even second century date is to suppose that the author is intentionally and quite successfully archaizing, but at that point one might well ask why we need that hypothesis at all, rather than the much easier one of saying that it looks like it fits into the pre-50 period because it indeed dates from that period. Absent compelling positive evidence that it must date to the later first or the second centuries, such an archaizing hypothesis seems to beg the question.
I can imagine only one really substantive objection to Robinson's dating, and that is the argument that the Epistle of James is responding to Paul's teaching on faith and works and thus must postdate his writings. Robinson notes some significant problems with this objection, however. Most notably, if the Epistle is responding to Paul, it fails to engage with or even be aware of the central issue in Paul's discussions of faith and works, namely Gentile inclusion in the nascent Christian communities. For the Epistle, this discussion remains entirely intra-Jewish. Robinson argues that Paul more likely represents a later development in the discussion regarding faith and works than the Epistle of James: what began as an intra-Jewish discussion has in Paul been translated from its initial context to the question of Jewish-Gentile relations. Frankly, this is more convincing than the idea that the Epistle of James is responding to Paul yet studiously ignoring the specific concern to which Paul was writing.