Markets got you down? Can't seem to keep up with your benchmark this year? Not sure you'll ever figure out this whole buy-low, sell-high thing that you keep hearing rumors about? Cheer up, you've got good company. Company anyway. Specifically, corporate America, whose overall market timing ability is much worse than that of a chimp who watches Art Cashin.
That's what McKinsey coldly concluded, noting that from 2004 to 2010, "Only 31 percent of the [S&P 500] companies earned a positive return from buying back shares—less than you would expect from a random throw of the dice." And not only did two out of three companies underperform putting cash under a pillow; three out of four underperformed dollar cost averaging:
Now, it is not exactly news that companies tend to screw up share repurchase programs from a strictly "don't lose lots of money" perspective. There is a simple and intuitive explanation that may even be correct. Most people buy stocks when they get a paycheck, or win the lottery, or think PE ratios are too low, or see a bullish engulfing candlestick pattern, or get a tip from an insider. Some of their buying is pro-cyclical ("ooh, stocks are up, I should buy some!"), some (more?) is counter-cyclical ("ooh, stocks are cheap, I should buy some!"), and some is uncorrelated ("ooh, I found $20 on the floor, I'll buy a round lot of BAC!").
Companies buy their own stocks when they are flush with cash, feel generally confident, and don't have any particularly good ideas for how to grow their businesses. You would not be crazy to expect that approach to correlate well with market tops. Indeed, McKinsey's study understates the under-performance, because it looks only at buys. Companies don't just buy stock when they're flush with cash and performing well - they also sell stock when they're out of cash and on the ropes. Buy high, sell low.
Never fear, though, McKinsey has a solution:
These findings suggest an easy fix: companies should give up trying to time the market. Long-term shareholders will be better off if management would simply forecast total excess cash and evenly distribute it each calendar quarter as “dividends” in the form of share repurchases. CFOs can approach such regular buybacks in two ways. First, they can repurchase shares as excess cash becomes available. This is the easiest approach and the one least likely to send adverse signals to investors around the potential for excess cash or cash shortfalls. It is probably right for most companies, even if it generates lower returns. Second, companies can evenly distribute similarly sized repurchases over time. For those willing to stand by their forecasts of future cash flows, this dividend-like approach will probably generate higher returns for shareholders.
Wait ... "repurchase shares as excess cash becomes available"? Isn't that the problem? Don't companies buy stock high when they have too much cash and sell low when they have too little?
We ran kind of a crude test just to see what McKinsey was smoking and, surprisingly, either not only does "buy whenever you have excess cash" work (i.e. outperform the S&P) over the 2004-2011 time period that McKinsey used, but - contrary to their guess - it actually outperforms just spending a flat amount each month:
The blue line represents your cost if you bought based on LTM free cash flow; the green line represents your cost if you bought the same dollar amount each quarter.* Cumulatively, $1 invested in line with cash flow from 2004 to 2011 would be worth $1.15 today (though it looked pretty bad in 2009), versus $1.12 for just dollar cost averaging - both way better than the typical S&P company actually did. It's a strange result, probably driven by the rather unusual fact that post-2008 corporate profits and cash flows have been pretty good, and companies have hoarded cash, while the S&P has been scraping where it was five years ago.
So, um, a tentative high five, McKinsey.
The savvy executive’s guide to buying back shares [McKinsey Quarterly]
* FCF ("free cash flow") average price = cumulative average price of buying the S&P 500 in an amount proportional to the previous twelve months' cash flow from operations less capex for the S&P 500 components as of today; DCA ("dollar cost averaging") average price = cumulative average price of buying the S&P 500 in an equal dollar amount each quarter; assumes buying as of the price on the first day of each calendar quarter. Zillions of caveats including dividends etc. and that this doesn't reflect individual component performance, just the aggregate for the index. Source: Capital IQ.