Over the last four years, since the buyback boom began, from the fourth quarter of 2004 through the third quarter of 2008, companies in the S.&P. 500 showed:
Reported earnings: $2.42 trillion
Stock buybacks: $1.73 trillion
Dividends: $0.91 trillion
Baruch has no problem with the dividends. You can’t fake dividends; the checks would bounce. But sadly he must take issue with the idea of the buybacks being a “return to shareholders” in the US.
Fact is, many buybacks in the US are “fake”.Yes, shares are bought back by companies in the open market, but many, even most, of them go to offsetting dilution from share options which vest in the quarter concerned. In many cases, the actual number of shares in issue goes down by much less than the number of shares bought in the market by the company. Often the number of shares out stays the same, and in the worst case the company can buy back shares worth billions, and see its net share count actually go up.
Take Qualcomm (not Baruch’s favourite stock, he has to admit) in 2006, which bought back roughly $2.7 billion worth of its shares, 63m-odd. This was roughly 4% of its market cap. The fully diluted share count in fact go up from an average of 1,686 billion shares in Q4 05 to 1,693 billion in Q4 06, an increase of 0.4%.
Basic rule: buybacks don’t count as shareholder return if they don’t reduce the share count. The point of a buyback is that it is a tax-efficient way of rewarding shareholders, especially if the share price is undervalued. Aside from adding buying pressure in the market that helps the shares to rise, it raises EPS by the %age of outstanding shares bought back, and if that company pays a dividend as well, at a given payout ratio raises the dividend per share and thus the dividend yield of the stock too. Baruch loves buybacks where appropriate. He actually prefers a good buyback to a dividend.
What Qualcomm, and much of corporate America has been doing in buying back stock, is in fact pointing out to shareholders how much they are being screwed by share options. Think about it. New US GAAP rules were set up partly to account for the true costs of share option issuance , which we can think of as a reverse buyback; share options dilute existing shareholders. Now US GAAP accounting is hard. Baruch doesn’t really understand it, and like everyone else sticks to pro forma accounting, which is what all the analysts publish their numbers in anyway. US GAAP is a big failure in this respect. But if a company is polite enough to buy back on the open market the shares it issues as options, hey presto, then we know what the actual cash equivalent cost to me as a shareholder the share option policy of that company is. Options are of course a very important part of how they pay their employees and directors.
In the case of Qualcomm, that is $2,658 million in 2006. Which wasn’t a bad year for them. They actually earned $2,800 million in 2006, pro forma, or $1.64/share. Obviously we have to deduct that $2,658m, leaving us with erm, uhh. Oh. Hang on is that right?
OK so post buyback QCOM earned practically nothing in 2006 in actual concrete returns to shareholders. The buyback was an unbelievably inefficient way of hiding dilution, going only to show how overvalued the stock was. What they should have done of course is not do a buyback but rather let shareholders get diluted by 5c instead, and have 4% less EPS growth. The stock would have been $5 cheaper, probably, and its executives moderately poorer from less valuable stock options. They would have done much better using the cash to pay a proper dividend instead.
Almost every big tech company I know in the US has committed this sin. Cisco is another infamous wastrel in this regard, although there the share count sometimes goes down (but not as much as it should).
So, Mr Norris! Let’s not count all US backbacks as shareholder returns please.
POST SCRIPT: actually I didn’t quite realise the point of Floyd Norris’ postwhen I wrote this. It was late at night, slightly insomniac blogging, so forgive me. I think he is probably trying to find out what happened to all of the industrial activity of the past 4 years; where did all the money go? Was it all fake? Well here’s the answer. The majority of it was pissed away buying overpriced shares to offset relatively minor dilution. Or, less charitably, it was paid to management in share options, the cost of which remain largely unaccounted for. In that sense, the earnings were indeed fake, if you assume the managers being paid partly in shares would not have done their jobs for only the cash component of their earnings.
Thinking about it more, it’s actually quite shocking. I was always of the opinion I should deduct the cash spent on buybacks to offset dilution from my valuation models. I did not, in the end, because no-one else did, and it would mean I would have a portfolio consisting only of non-US stocks — a quite significant benchmark risk; it cut the fair value of the US names by about 2/3rds. There’s an argument to be made that that’s where they should be going.