This, to me, is the ultimate measure of effectiveness of an SQA group. To convert this into their efficiency, you'd have to take into account how much effort went into finding the bugs internally.
I suspect this varies widely from company to company. I do remember reading somewhere many years ago that someone had estimated that only 1/2 of the bugs that ever existed in a product were found before it shipped. I think that was focused on commercial, shrink-wrap products.
The funny thing is, there is not a good correlation between these numbers and the customers' perception of the quality of the products produced by certain companies. I remember one company I worked for that had almost as many bugs coming in from the field as from SQA, yet the customers seemed to feel that the product was fairly solid. Another place I worked had almost no bugs reported from the field, yet customers were always complaining that the products stank. Perhaps this is a function of the support group and their ability to characterize every customer call in the defect tracking system (the first company had a really good support system).