Zeev,
I've tried to keep out of this debate, but I just can't take any more. So here are the "correct" answers to some of the issues raised <g>. On the issue of share buybacks vs. dividends, you are correct and Culmus is wrong. Miller and Modigliani showed a long time ago that dividends and buybacks are economically equivalent in the absence of taxes. Culmus makes the point that INTC's share price would be higher by buying stock than by issuing dividends, which is true enough... but only because the number of shares would be decreased. The capitalization would have been the same in either case (fewer shares times higher price), and stockholders would have been no better or worse off.
You think MRK would have been better off buying some smaller companies than paying dividends? Maybe, maybe not. Don't forget, the shareholders that received those dividends could have reinvested them in the same smaller biotech companies...they don't need MRK to do this for them. So it really comes down to whether you think MRK management could and would have made better investment decisions than their shareholders. Frankly, corporate America's record of adding value through acquisitions has been pretty dismal. On the other hand, a company like Berkshire Hathaway has done a good job in this regard. Still, I'd bet against you on this question, just based on the history of corporate acquisitions.
On your emphasis on tangible book value, I'd generally disagree. Depending on the company, BV is somewhere between interesting and totally irrelevant. There is a large discrepancy between BV and market value for companies like DELL and CSCO. The value created is the difference between the market value (not book) and net contributed capital. Somebody who bought DELL at the ipo and sold today has real wealth, not theoretical. The same point applies to your problem with dilution of BV though share buybacks, which does not trouble me at all. BV is an accounting construct. The share price is what it is.
Okay, I'll shut up now.
AI
Market Data 
Markets 
