The only reason not to buy back shares if they're trading below tangible book value [1] is because you have investments available that will yield greater returns, or do not have sufficient cash to do so.
[1] Factoring in a margin of safety, of course.
His point is that companies that can't find better ways to use capital than a buy back aren't innovative enough to succeed in the long term
Stock buybacks are very popular with some investors as a way for a company to transfer value from the company back to the shareholders. It is like a dividend, except it is taxed as a capital gain, not ordinary income.
He's approaching it from the opposite direction; ie returning value to investors who are selling their stock. But if the company is truly worth more per share, in terms of tangible book value, than the price the shares are trading at then anyone who is selling up is getting a bad deal. These people aren't getting value. It's the people who retain their stock who are getting value because they're managing to buy slices of a company worth $(X+Y) for only $X.
A simplified example:
FictionCorp has $200 million in the bank and 100 million shares outstanding. For some reason, the shares are trading at $1, despite the $2/share sitting in the bank.
If FictionCorp manages to go and buy back half its stock at $1/share, it now has $150 million in the bank but only 50 million outstanding shares. Each share now represents $3 worth of cash in FictionCorp's bank account.
Hence, the remaining stockholders are better off because they've managed to buy a larger portion of FictionCorp at a discount to its actual value.
Go buy shares at $1.01 during the buyback (people will sell to you rather than FictionCorp, since you are the best bidder), then sell them at $3.00 after the buyback is finished. This is a moneymaking opportunity whenever the share price is below $2.00.
Come back when you are rich and I'll believe you found an inefficiency in the market. Until then, I think you are wrong.
If you believe markets are always efficient[2] then I ask you to explain the tech crash in 2001. The NASDAQ lost 78% of its value over two and a half years and is still down ~60% from its peak.
[1] Of course if you don't have control over the management and they're spending profligately enough to eliminate the advantage you wouldn't invest. Fred Wilson was the Chairman, however, so that was a non-issue.
[2] They are mostly efficient, but at times human psychology causes them to be over or undervalued, ie during bubbles or after crashes.
Buybacks work in theory, but they depepnd on a lot of factors.
Essentially what you are trying to do is shrink the existing share count, so that EPS goes up which presumably drives your share price up in the longer term.
E.g.:
$10 in net income / 10 shares = $1 EPS. Lets say right now you are trading at a 6x multiple, the market is nervous about your prospects. You think your shares are undervalued and that you are probably worth 10x (perhaps your industry average).
So you buyback 20% of your shares, existing share count decreases from 10 to 8. EPS goes up to $1.25 per share.
But, one of the uncertainties behind a buyback is the fact that you never know when or how much management will actually buyback. They will sometimes announce a plan that would buyback 20% of shares, but only do that over maybe 10 years. Thats why your bit about timing the buyback just doesn't work.
Let's say somehow you do manage to buyback the 20% in just one year (the market hates your company). So with an EPS of $1.25, if the market keeps giving you that 6x multiple, in theory your shares should be worth at least $7.50, not the $6 before. Moreover, if the buyback signals enough confidence in the company, maybe the market will realize it and you will return to the 10x multiple you had before, so your shares go to $12.50
But see nobody can time when the market works or when the market will catch on to the value of your company. Benjamin Graham said that is one of the great mysteries of the stock market when he testified before congress.
It is during that lag time when the share price does not move that investor confidence can waiver and maybe they will sell -- ultimately not realizing the value that eventually comes from the buyback and reduced share count.
I don't know where you get that idea. I work for a hedge fund, which would be really dumb if I believed markets were efficient. That would mean my expected bonus is $0 [1].
When a rich person tells me "markets were inefficient, and now I'm rich", I often believe them. When a person tells me "look, a whole bunch of idiots are throwing their money away, but I'm not going to take it off their hands", I'm a little skeptical. And when someone says "look, the markets are inefficient in a manner that is extremely obvious and easy to exploit", I'm also a bit skeptical, since most such inefficiencies have already been arbed away.
Lastly, my trading strategy doesn't require the buyback to complete. All it requires is that philk's claim actually occurs: that the buyback transfers money from sellers to shareholders. Using his numbers and my proposed buy price of $1.01, the moment the buyback has transferred at least $0.02/share from sellers to shareholders, the strategy has become profitable.
[1] Not strictly true, since a bonus behaves like a call option on profits, but you get the picture.