No abstract

IntroductionIn this study, we examine the relation between chief executive officer (CEO) ability and management earnings forecasts. Trueman (1986) theorizes that, because a manager's ability to identify changes in his ⁄ her firm's underlying economics is value relevant, a manager with equity-based incentives will voluntarily issue an earnings forecast as a signal to the market that the manager has identified such changes. Thus, the forecast release will lead to a higher firm value at the end of the period than if the forecast had not been issued, thereby enhancing the manager's equity-based wealth. This implies that the forecast contains useful information about CEO ability, incremental to the earnings information in the forecast. Specifically, we examine the relation between CEO ability and both the likelihood and frequency of management earnings forecast issuance. We also examine the relation between CEO ability and both forecast accuracy and the market responsiveness to the news in the forecast.CEO ability is difficult to observe directly. Consequently, a nontrivial issue with testing Trueman's 1986 theory is identifying empirical proxies for CEO ability. This is a plausible reason why to this point there has been little empirical work done to test Trueman's theory. We seek to fill this gap in the literature with our study. To execute our tests, we adopt three recently introduced proxies for CEO ability. Our first proxy for CEO ability is based on the number of press citations for a CEO over a prior five-year period. This measure is consistent with that used in Milbourn 2003, Rajgopal, Shevlin, and Zamora 2006, and Francis, Huang, Rajgopal, and Zang 2008. The intuition behind the press citation measure is that more talented CEOs tend to be cited more in the press. Our second measure of CEO ability comes from Demerjian, Lev, and McVay 2009, who use data envelope analysis (DEA) to create a measure of manager-specific efficiency (i.e., ability). DEA has been utilized in other settings, such as research investigating the relation between managerial ability and bankruptcy (e.g., Barr and Siems 1997;Leverty and Grace 2009). For our third measure of CEO ability, we follow Rajgopal et al. 2006 and use industry-adjusted return on assets during the prior three years of a particular CEO's tenure.Using a large sample of firms over the period 1995-2005, we find that the likelihood of management earnings forecast issuance increases in CEO ability. This result is consistent with Trueman's 1986 theory that high-ability managers are more likely than low-ability * Accepted by Patricia O'Brien. We thank two anonymous reviewers,

In a computer system, the integrity of lo~ler layers is typically treated as axiomatic by higher layers. Under the presumption that the hardware comprising the machine (the lowest layer) is valid, integrih of a layer can be guaranteed if and only~: (1) the integrity crf the lower layers is checked, and (2) IntroductionSystems are organized as layers to limit complexity. A common layering principle is the use of levels of abstraction to mark layer boundmies. A compuier system is organized in a series of levels of abstraction, each of which defines a "'virtualmachine" upon which higher levels of abstraction are constructed. Each of the virtual machines presupposes that it is operating in an environment where the abstractions of underlying layers can be treated as axiomatic. When *Arbatrgh is also with the U.S. I@wtment of Defense. fsnuttr and Far&r's work is snpp(rted by DARPA under Contracts #DABT63-95-C-O073, #N66C01-%-C-852.and #MDA972-95-l - CIO13 with additionalSUppOrt from Hewlett-Packtud and ln~cl~orlwrations these suppositions are true, the system is said to possess integrity. Wkhout integrity, no system can be made secure. Thus, any system is only as secure as the foundation upon which it is built. For example, a number of attempts' were made in the 1960s and 1970s to produce secure computing systems, using a secure operating system environment as a basis [24]. An essential presumption of the security arguments for these designs was that system layers underpinning the operating system, whether h,ardware, firtnw,are, or both, are trusied. We find it surprising, given the great attention paid to operating system security [161[9] that so little attention has been paid to the underpinnings required for secure operation, e.g., a secure bootstrapping phase for these operating systems.Without such a secure bootstrap the operating system kernel cannot be trusted since it is invoked by an untrusted process. Designers of trusted systems often avoid this problem by including he boot components in the trusted computing base (TCB) [7]. That is, the bootstrap steps are explicitly trusted. We believe that this provides a false sense of security to the users of the operating system, and more important, is unnecessary. AEGISWe have designed AEGIS, a secure bootstrap process. AEGIS increases the security of the boot process by ensuring the integrity of bootstrap code. It does this by constructing a chain of integrity checks, beginning at power-on and continuing until the final transfer of control from the bootstrap components to the operating system itself. The integrity checks comp,we a computed cryptographic hash value with a stored digitiaJsignature associated with each component.The AEGIS ,at-chitecture includes a recovery mechanism for repairing integrity failures which protects against some classes of denial of service attacks. From the smrt, AEGIS 65

F m i n ( x ) = 1 -exp [-exp [a1(~ -~1 ) l I where 1 n F(u1) =andSince, in general, the underlying initial distribution is unknown, the parameters u,, a,, u1, and a1 have to be estimated from sample values. This usually proceeds as follows. A total of k = nN samples from a given distribution are taken and divided into 'N groups of n samples per group.If it is desired to estimate u, and a,, in each of the N groups the largest of the n samples is chosen. From these N largest values, it can be shown that a variety of procedures exist from which u, and a, can be estimated. A method for obtaining maximun-likelihood estimates of u, and a, is described in [9] , along with various other estimation techniques.If it is desired t o estimate u1 and a1, a similar procedure using minimum instead of maximum values can be performed. REFERENCESD. J. Gooding, "Performance monitor techniques for digital receivers based on extrapolation of error rate," ZEEE Trans.

In this study, I examine the association between the credibility of the financial reporting system and the quality of governance mechanisms. I use a sample of 87 firms identified by the SEC as fraudulently manipulating their financial statements. Consistent with prior research, results indicate that fraud firms have poor governance relative to a control sample in the year prior to fraud detection. Specifically, fraud firms have fewer numbers and percentages of outside board members, fewer audit committee meetings, fewer financial experts on the audit committee, a smaller percentage of Big 4 auditing firms, and a higher percentage of CEOs who are also chairmen of the board of directors. However, the results indicate that fraud firms take actions to improve their governance, and three years after fraud detection these firms have governance characteristics similar to the control firms in terms of the numbers and percentages of outside members on the board, but exceed the control firms in the number of audit committee meetings. I also investigate whether the improved governance influences informed capital market participants. The results indicate that analyst following and institutional holdings do not increase in fraud firms, suggesting that credibility was still a problem for these firms. However, the results also indicate that firms that take actions to improve governance have superior stock price performance, even after controlling for earnings performance. This suggests that investors appear to value governance improvements.