Abstract
The purpose of this note is to clear up several misconceptions presented by Saghafi and Attaran (1991) (S-A), in their reply to my comment (O'Neill, 1991a). The most egregious misconception being that S-A failed to recognize the main point of my comment: the entropy index values reported by Attaran and Saghafi (1988) (A-S) are erroneous becouse they are based on incorrectly computed firm market share values. The problem being that A-S compute share values using the sales (or assets or equity) of the Fortune 500 firms in the denominator of their share figures rather than using the total manufacturing sector saless (or assets or equity). This obviously overstates each firm's market share. As stated in O'Neill (1991a) the effect this has on the entropy index cannot be determined aa priori, it depends on the data sample. In the entirety of their rely, not once do S-A refute this claim. A second, perhaps more important, point is that when the entropy index is computed using corrected market share values, the trend of the index over time is exactly the opposite of that which A-S originally reported. Aa-Ss present entropy index values which are declining betwweeen 1970-84, indicating rising concentration in manufacturing during this period. The corrected index values show that the index has been rising throughout this time period. Corrected index values for 1967-87 are presented in O'Neill (1991b) and uupdated below. Further-more, the 'new evidence' presented in S-A is also misleading (as is discussed below).
Let me now turn to specific details in S-A, as they make several erroneous interretations of my comments.