Pacwa says when Luther translated the Bible, he was simply using an already existing German translation of the Bible which he worked with and altered. One can't help then but conclude according to Pacwa, Luther didn't really translate the Bible, but rather revised a pre-existing translation. Attaching this to Luther's rendering of Romans 3:28, Pacwa basically indicts Luther of simply putting a word in an already existing translation. It was an alteration of Romans 3, not a translation.
There is indeed some interesting facts here. Philip Schaff addresses this controversy. After discussing pre-Reformation German Translations of the Bible, Schaff states:
Luther could not be ignorant of this mediaeval version. He made judicious use of it, as he did also of old German and Latin hymns. Without such aid he could hardly have finished his New Testament in the short space of three months.(8) But this fact does not diminish his merit in the least; for his version was made from the original Hebrew and Greek, and was so far superior in every respect that the older version entirely disappeared. It is to all intents a new work.
(8) Luther's use of the older German version was formerly ignored or denied, but has been proved by Professor Krafft of Bonn (1883). He adds, however, very justly (l.c. p. 19): "Es gereicht Luther zum grössten Verdienst, dass er auf den griechischen Grundtext zurückgegangen, den deutschen Wortschatz zunächst im N. T. wesentlich berichtigt, dann aber auch mit seiner Genialität bedeutend vermehrt hat."
So what's going on here? Both Pacwa and Schaff make the same point, but slightly spun different ways. For Pacwa, Luther simply altered an already existing translation. For Schaff, Luther did use the original Greek and Hebrew, but did use the older translation in order to finish his translation quicker.
This questioned was tackled head on by Johann Michael Reu, who was one of the foremost experts on the production of Luther's Bible. In his book Thirty Five Years of Luther Research. He explains it was actually a Lutheran who made the first helpful evaluation of pre-Reformation Bibles, W. Walther. From this study, Krafft contended that Luther had indeed used an earlier Bible as a translation source. This tidbit was then picked up by others, like Vedder, and expressed as:
"This version was certainly in the possession of Luther, and was as certainly used by him in the preparation of his version. This fact, once entirely unsuspected, and then hotly denied, has been proved to a demonstration by the "deadly parallel." It appears from a verse-by verse comparison that this old German Bible was in fact so industriously used by Luther, that the only accurate description of Luther's version is to call it a careful revision of the older text. Just as the English Bible is the result of successive revisions, from the days of Wiclif to our own, so that our text has a demonstrable historic continuity, so the German Bible is the product of revision. This is not to detract in the least from the glory of Luther or to diminish the value of his version—it is merely to define with accuracy what he accomplished, and to distinguish his real achievement from the semi-legendary tales of Lutheran literature."
"It would be difficult in any case to believe that a complete translation of the entire New Testament could have been made by a man of Luther's limited attainment in Greek, and with the imperfect apparatus that he possessed in the short space of ten weeks. . . . Any minister to-day who has had a Greek course of a college and seminary is a far better scholar than Luther. Let such a man, if he thinks Luther's achievement possible, attempt the accurate translation of a single chapter of the New Testament—such a translation as he would be willing to print under his own name—and multiply the time consumed by the 260 chapters. He will be speedily convinced that the feat attributed to Luther is an impossible one."[source]
"Walther shows how just in all those places, where the use of the mediaeval Bible through Luther must have shown itself, granted that Luther used it at all,—for example, in difficult passages,—that just there entirely different translations are to be found, different not only as to the words used, but also as to the method of translation in respect to style as well as to syntax. Parallels only show themselves there where the renderings—especially in the historical books—might, because of their nature, be alike, without being copied. If Luther really was acquainted with the Bible of the Middle Ages, he did not use it. During the first phase of his translation work, and the one that gave the work its characteristics, he was not acquainted with it, as we can state with a reasonable degree of definiteness. Only later he became acquainted with it, and then, as we can see now, in his revisions and corrections he occasionally supplanted his own word with one from it."
"And just this we pronounce childish argumentation [Vedder]. We could call attention to the fact that R. P. Olivetan completed his French translation of the entire Bible, printed 1536, in one year; that Luther finished his writing against Sylvester Prierias, that in Walch's edition fills 80 columns, in two days; that Luther was in fact a linguistic genius; that an educated man in the thirties acquires a dead language much faster and more thoroughly than a youth from 16 to 20, and this all the more, the dearer and more valuable, yes, even decisive for his whole life, the contents of a book written in that language is to him; that Luther since 1519 had been a careful reader of Homer, writing many marginal notes into the copy which Melanchthon had presented to him (this copy is extant at London, cf. Pr. Smith, Notes from English libraries, Zeitschr. f. Kirchengesch. 32, pp. 111-115; compare also: O. G. Schmidt, Luther's Bekanntschaft mit den alten Klassikern, 1883). We also could emphasize the important fact that Luther for more than ten years was well versed in the contents of the New Testament through the Vulgata. But aside from this we would like to ask Vedder whether he has forgotten that Luther, as can be proved, since 1516 used the Greek original in the preparation of his lectures, and certainly not seldom also when he prepared his sermons, and that it more and more became the foundation for his whole theological work; that he, before his stay at the Wartburg, had treated the Epistle to the Romans, Hebrews, Galatians, perhaps also the Epistle to Titus and the first to the Corinthians in lectures, the Epistle to the Galatians beside this also in a voluminous commentary; in short, that Luther lived and moved in the New Testament, and, finally, that the printing of his translation had not begun for two months after his return from the Wartburg, and that it was not completed before six months had passed? During this time he, together with Melanchthon and other occasional helpers, once more revised the whole in a most painstaking manner."
Does Reu set the record straight? I'd have to see what analysis has been done since Reu. Reu does admit Luther used it, but not to the extent expressed by Vedder and Krafft.