I appreciate the opportunity I’ve been given to post on this Harvard blog.
Appropriately enough, I’m going to start with a response to the work highlighted on this blog of Bebchuk, Grinstein & Peyer (BGP) in Lucky Directors and Lucky CEOs. As summarized in the Financial Times op-ed reproduced on this blog, BGP say their work “suggests [opportunistic timing of option grants] deserves all the attention it has been getting and more.” They show that backdating has produced significant gains, been widespread across industries, occurred mainly when the pay-offs were high (i.e., with larger gaps between lowest and median prices in the grant month), was correlated with evidence of CEO influence (in firms lacking a majority of independent directors and where the CEOs had longer tenure), and occurred where CEOs had more total compensation from other sources (suggesting, they say, that it did not substitute for other compensation). The director study shows that backdating was prevalent for outside directors. The authors conclude that “[t]he patterns we have studied reflect persistent, widespread and systematic governance problems” that “deserve investors’ full attention.”
Here I take issue not with the authors’ basic data but with their conclusions about what the practice says about corporate governance generally.
The important question is whether the backdating could be considered a form of theft. If it was, then this, indeed, indicates a serious governance problem. Moreover, the fact that the practice was evidently correlated with weak governance constraints (more entrenchment, fewer outside directors) suggests we should correct those problems.
However, notwithstanding BGP’s data, there are several reasons to think that backdating was much more benign:
1. As Holman Jenkins and others have pointed out, backdating simplifies negotiations between executives and their firms, especially in firms with volatile stock prices (one of the correlations BGP find). The firm decides how much it wants to pay the executive and then picks an option strike price that produces that compensation, rather than negotiating the pay against a mere expectation of what the price will be on the date of the grant.
2. Of course the firm could have done the same thing by simply fixing the strike price rather than the date. Backdating reduced the company’s accounting expense and the recipient’s taxes (and the company’s expense deduction). To this extent there may have been dishonesty. But there is so far no evidence from BGP or others that this increased executives’ compensation vs. non-backdated awards. The fact that backdating wasn’t correlated with lower compensation from other sources does not show that total compensation would have been lower without backdating. Indeed, the simplification rationale (in 1) suggests the backdating was used to reach a target figure, which would have remained the same in the absence of backdating.
3. This conclusion is further supported by the observation of Stanford’s Joe Grundfest that accurately reporting the options as in-the-money would have been “hellishly difficult.” Accounting rules encouraged firms to use a shortcut. Again, this is not contradicted by the fact that other compensation wasn’t lower in backdating firms.
4. Even if backdating increased pay, the increase was less than would be indicated by BGP’s differences between manipulated and non-manipulated option prices. It is important to keep in mind that, under Black-Scholes, the discrepancy in value is considerably less than the difference in option prices.
5. The correlation with weaker governance does not itself show fraud. Long-serving CEOs are likely to have been the most successful, and therefore the highest-paid. CEO success, in turn, has been shown to lead to management-selected boards and weaker governance. See, e.g., Hermalin and Weisbach, Endogenously Chosen Boards of Directors and Their Monitoring of the CEO. We would expect more successful CEOs to be paid more, and to have more clout in nailing down that pay, including through backdating.
6. There is no indication that the backdating itself fooled the market. The firms, after all, disclosed the option price and the number of options. This is the key data, and there is every reason to believe that, as the New Yorker’s James Surowiecki points out, “accounting legerdemain doesn’t fool the market.” Thus, despite the machinations, there is no justification for calling this “stealth compensation.” Narayanan, Schipani and Seyhun’s much-publicized study concludes that revelation of backdating resulted in an average loss of 8% or $500 million dollars in stock value per firm. But the study does not employ industry and other filters to control for other events affecting these firms. The significance of this omission depends on whether you accept David Walker’s conclusion that this was a high-tech phenomenon, BGP’s conclusion that it was widespread across industries, or whether there was an additional correlation none of the studies have picked up. More seriously, the drop in price probably reflects the concern over punishment, particularly by the press, which has had a field day with backdating. Note also that, as the NYT has reported, plaintiffs’ lawyers have been bringing derivative suits in backdating cases because they haven’t been able to find market damages. Even Melvyn Weiss was moved to comment that “a lot of these stocks aren’t reacting with big drops in price. Or, if they drop initially, they come back over a short period of time.” If there really were serious stock-price effects, nobody is more motivated than Mel Weiss to find them.
7. Even if there was accounting fraud, the discussion above suggests that this does not signify massive governance failure. The very widespread nature of the problem found by BGP and others itself suggests a prevalent belief that this practice was benign. As Holman Jenkins has said, “embroiled are not just a few bad apples . . . but Apple, Pixar, Microsoft, Juniper Networks and, if Prof. Lie is to be believed, 29.2% of all listed companies. In such circumstances, ‘the CEO is a crook’ is an explanation that quickly loses power.” In assessing the crookedness of the conduct, we should also consider some other facts bearing on what the executives thought they were doing. The accounting rules, at best, were a long way from moral guidelines. Controversy still rages over whether stock options should be expensed at all. See Kip Hagopian, Point of View:Expensing Employee Stock Options Is Improper Accounting, whose conclusions have been endorsed by Milton Friedman, Harry Markowitz, Paul O’Neill, and George Schultz, among others. And if stock options are appropriately expenses, the same should be true whether they are in, at, or out of the money. So we can see why firms might have given rules that distinguished among these categories somewhat less moral authority than the Ten Commandments.
It is also significant, as Holman Jenkins has pointed out, that the firms were hardly covering their tracks. Rather than fudging the option timing in a way that would have provided murky support for manipulation, they were blatantly taking the date with the lowest price in the relevant period. All of this was disclosed for the convenience of the researchers who came along and discovered it. Let me emphasize that none of this is to excuse the practice, but only to question its lessons for corporate governance generally.
8. Given that the practice was wrong, we should do something about it. But what? BGP suggests a lot. But the factors discussed above indicate that the problem was not as malignant as press reports imply, and the fact that companies themselves are dealing with it by firing implicated employees, indicate that more drastic steps such as criminalization of the conduct are inappropriate. The bottom line is that the BGP papers, though helpful, are no more than what should be the beginning of studying the institutional background of backdating. At this point we do not have a smoking gun indicating a serious governance problem. BGP eliminate some explanations that few believed anyway, such as that “independent” compensation committees would produce better compensation. Before using the practices as the basis for overhauling corporate governance, we need to understand more about the institutions that produced them.
A final note: one might suppose that I’m tempted to minimize the backdating problem in accordance with my own Panglossian view of corporate governance. In fact, my view of the current state of corporate governance is darker even than Professor Bebchuk’s. While Professor Bebchuk believes that corporate governance can be repaired by fixing some voting rules, I have repeatedly suggested a more drastic revamping of corporate governance, and indeed substitution of a different “partnership-type” framework. See my Accountability and Responsibility in Corporate Governance. If I were inclined to look for support for my theory, the widespread nature of backdating across industries, and its resistance even to governance fixes like independent compensation committees, would be just what I was looking for.