The American Historical Association's Blog reports that:
This completion rate is significantly lower than our own estimates, which are based on data submitted from the history departments annually. In 2006, history PhD programs estimated that an average of 59 percent of the students who matriculated into their program in 1996 had completed the degree. But one of the issues we confronted in our survey was precisely how to determine when a student actually entered the doctoral program, given that some enter as Master’s students or some other form of transitional status....
Unfortunately, the report does not offer specific information on attrition from history PhD programs. But for the humanities in general, they report that slightly more than 15 percent of doctoral students gave up by their third year in the program, and more than a third gave up by year 10. Almost 19 percent of humanities doctoral students were still counted on the books as continuing toward the degree in their tenth year.
The discrepancy in reporting is important to focus on. The AHA Blog goes on to conclude that the report gives "history a lower 10-year completion rate than almost all other disciplines."
Another way to look at the very same tables is to find that while it takes longer to complete a PhD in the humanities than mathematics or science (we knew that), at the end of the day, the 10 year rate for history is not far off from many fields, including economics, which has a 10 completion rate of just over 50%. The table that is supposed to show that history has a lower completion rate than "almost all other disciplines" actually shows that there is almost no difference between history and the overall category of "humanities." This suggests that if steps need to be taken to help graduate students complete the PhD more quickly, it is not necessarily a history-specific problem, but a difficulty that affects many parts of the academy.
The bottom line: if you want a quick PhD, the data is clear: be a civil engineer. It you're in econ, history, anthro or literature, you're in for a long haul, and it is in everyone's interest to determine ways to make progress toward the PhD, across fields, an education, and not a career.
UPDATE: PhDinHistory follows up with more helpful data, here, noting that only sixteen PhD granting history departments provided the date for the Council of Graduate Schools report, representing only 10% of North American history doctoral programs. A rather important limitation of the study!