This isn't quite right. Look at it this way:

There is a company that has the following assets

  1) $100 in cash
  2) $100 in other assets

The company spends its $100 in cash to buy back 100 shares. So now its asset list is:

  1) $100 in other assets

So why do a buyback if it doesn't raise the share price? Two reasons:

  1) it returns cash to the shareholders who want it (those that sell the stock)
  2) future gains in value will now accrue to fewer shares of outstanding stock

(Just wanted to clarify that the benefit is independent of the ticker price subsequently increasing)

Therefore, in your above example if the shares were actually being sold at $2 each, then the company is doing a disservice to the investors by buying the stock back.

However, if the stock is under pressure (below its “fair” price), say 50 cents in the above example, then by buying the stock back, you are creating true value for your investors.

Problem of course is that rarely do people running the company think this way.

Which means that your company just gambles its money at the stock market by betting on its own price.

One reason an investor might buy shares is to try to take over the company, i.e., end up with all of its shares. Well, it is well known that the people executing a takeover often depend on getting access to the target's cash and other easily-measured (and consequently easily-sold) assets to help pay for the takeover. Specifically, if the company has $100 in cash, the takeover artist can afford to pay approximately $100 more for the company than if it has $0 in cash. (Specifically, if the target has $100 in cash, the takeover artist can go to one of his friends and says, "Loan me $100 and I'll pay you back with the cash held by the company I am planning on buying.) So since market cap is approximately what it would cost to buy the whole company, it makes sense that a reduction in cash would cause a reduction in market cap of approximately the same amount.

Note that even investors who would never contemplate taking over a company can profit by selling their shares to someone who would and does, so they have the same incentive for valuing the company's cash as the takeover artist does: namely, the takeover artist can afford pay more for a company with cash, so the price paid by the takeover artist to the investor will be commensurately higher.

Now the fact that management decided to buy the company's shares rather than deploy the cash in some other way does send a signal to the market that management believes the company's stock is good value with the result that the price might end up at $103 instead of $100. We might call that a "second-order effect" of the buyback. Another second-order effect is that fact that if the market knows a buyback is in progess, they'll tend to hold out for a higher price. (In other words, a big purchase tends to drive up the price; the company will spread the buying out over time to try to limit that effect.) But these second-order effect are limited in magnitude; they are very unlikely to cause the price to end up at, e.g., $200.

It's a form of returning capital to shareholders popular in US companies which by-passes US double-taxation of dividends and is often preferred by shareholders as they're able to choose when to realize their capital gains at their optimal convenience and at a cheaper tax rate than regular income.

But they're typically performed when the company believes their share prices are undervalued and would like to acquire them at their perceived discount.

A lot of the valuation of public companies are speculative based on their future potential earnings so a loss of unused capital shouldn't affect it too much, it's also a signal that the company believes their shares are currently undervalued, so they believe their Buy Backs are acquiring them at a discount.

What if a company does buybacks on a predictable schedule over years?

It seems to me that if the pushing up is as predictable as you say it should be trivial to buy a bunch of shares right before the increase and sell them right after for an easy gain. You should get on that!

It depends on the size of the BuyBack and whether they're getting them at a discount, but it could go up at announcement as it could encourage other investors to buy shares early in expectation that the Buy Backs will push the price up, it will go up (depended on the size of Buy Back/Market Cap) during execution due to the increased demand.

> What if a company does buybacks on a predictable schedule over years?

They're only going to perform Buy Backs when they think they're at a low valuation, when they have a high valuation they'll prefer to give out dividends or employ other effective use of capital like M&A to manufacture growth.

> It seems to me that if the pushing up is as predictable as you say it should be trivial to buy a bunch of shares right before the increase and sell them right after for an easy gain. You should get on that!

You don't get to know exactly when Buy Backs occur, they'll announce that they intend to perform purchase a certain amount of Buy Backs over a specified period and in the next Financial results they'll disclose how many Buy Backs they purchased during the period.

Also your maths is slightly off. If the company uses $1bn to buy back stock, the company is now $1bn poorer. So whilst EPS etc will increase the assets per share may well decrease so I wouldn't necessarily expect the value of the outstanding stock to increase by $1 billion.

[1] http://www.barelkarsan.com/2010/12/buybacks-in-style-at-wron...

Right. Before, the shareholders have shares. Afterwards, they have shares and money. The total value has not changed at the moment of the operation (ignoring taxes, assuming infinity liquidity, that the information about the transaction was incorporated already in prices, etc.).

> It's effectively a dividend, by raising the stock price by the same amount they use to buy back stock.

It's effectively a dividend, but it's different in that the share price remains constant for a buyback and goes down for a dividend (again, in the ideal case).

With a dividend: Before, the shareholders have X shares. Afterwards, they have money and X shares (which have a price lower than before, to keep the total constant).

With a buyback: Before, the shareholders have X shares. Afterwards, they have money and Y shares (which have the same price as before, but there are less of them to keep the total constant).

> Like if they use $1 billion to buy back stock, the aggregate value of the outstanding stock should go up by $1 billion.

Nonsense. It's the other way around: the aggregate value of the outstanding stock should go down by $1 billion.

Where "should" applies only in the ideal case, sure, but there is no way at all to construct an argument where it "should go up by $1 billion".

To buyback stock the company has to make an offer that enough investors are willing to sell. Once investors start to do so at these higher prices, the stock price goes up.

To the low information investor, this looks to them like the stock is worth even more than it was before, because it made money for them (on paper, at least).

And what about index funds? Regardless of the number of buybacks, the index fund isn't going to sell. This would also cause the stock price to rise.

This idea that a stock buyback doesn't affect price only works in a market where everyone has perfect information. In real life this isn't the case.

It’s kind of strange that people hate on buybacks all the time (people like Elizabeth Warren), but no one gives a rat’s ass about dividends.

Er, no. That's not how it works. Companies can create new shares out of nothing anytime they want.

But the real issue is that shareholders will be informed when new shares are issued and no one likes dilution. I’m not 100% sure, but I believe existing shares held by the company can be sold without notification until a quarterly report.

Companies have a number of authorized shares and typically need shareholder approval to increase it. Sure, they typically set the number of authorized shares high enough to run into the problem with any frequency, they can't "create new shares out of nothing anytime they want."

One follows from the other. It's because it allows a tax dodge.

I preferer the money - id make an exception for discount control on investment trusts though.

There are I think better and more transparent ways of returning $ to investors.

I think you are confused. A buyback returns money to shareholders because the company has literally purchased the shares in exchange for money. Any other effects of a stock buyback are secondary. There's nothing vague about the promises, you're guaranteed to receive the money, and there's nothing to ignore with regard to the time value of money, because the shareholders immediately get money. Unless they decided not to sell, in which case, who cares? If they thought they would be better off taking the money they would have sold.

Your confusing a tender offer at over par with a generic buyback.