From Daniel on IBM (from my fool board)
Speaking of annoying, it looks their average net cost basis per share for stock buybacks in 2001 was $185.33. (Common shares outstanding went down by 19,705,455. Net common stock transactions were -$3,652,000,000. Note: I think that was before $502,000,000 worth of tax benefits from employee compensation in the form of stock and stock options, after which the average net cost basis per share would've been $159.85. But, the tax benefit wouldn't happen on the repurchases themselves.)
Of course, the market price was under $120 for most of the year, and sure enough, the book average cost basis per share repurchased in 2001 was only $106.51. (Number of shares of treasury stock went up by 59,278,078. Total cost of treasury stock went up by $6,314,000,000. Note: the employee benefits trust was eliminated during that year, and the 20,000,000 shares were transferred to treasury stock at a repriced cost of $2,255,000,000. Excluding them, the number of shares went up by 39,278,078 and the total cost went up by $4,059,000,000 for a more accurate book average cost basis per share repurchased in 2001 of $103.34.)
AFAIK, the only possible reason for that massive discrepency between effective cost basis and book cost basis was the laundering of employee compensation through the stock market via ESO's, to keep it off the income statement. Since they don't show all the details (these shares sold at this price, those shares purchased at that price), the net change in shares on the market over the year handily disguises significant turnover, as shares were issued, sold to employees via low strike price options, and repurchased (or pre-purchased) in the same year at full market value (1,253,156 shares just for Gerstner in 2001).
However, using the previous calculations, I believe there's a way to estimate how much money was lost by buying high and selling low. Intuitively, the difference between the effective average repurchase cost basis and book average repurchase cost basis, divided by the effective average repurchase cost basis, times the net common stock transactions, should approximate how much of the repurchase cash burn was churn. (($185.33 - $103.34) / $185.33) * $3,652,000,000 = somewhere between $1,615,000,000 and $1,616,000,000 lost on churn between January 1 and December 31.
That's not trivial. In fact, it's about 21% of 2001 net income. And, now that they're in the habit of churn, you should consider it a normal cost of doing business for them, just another (albeit bizarre) method of transferring money to employees...which means that 2001 net income applicable to common (which already took the tax break into account) was really only about $6,098,000,000.
Well, that was fun to figure out. ;) I hope someone finds it helpful.
Cheers,
Daniel Dauenhauer |
