Bharat Sarath

Uncertain Litigation and Liability Insurance

Legal penalties and liability insurance seem to have counteracting effects on the incentives of a potential injurer to take due care. However, if legal penalties are set efficiently and implemented perfectly, unrestricted access to insurance can be optimal. In contract, if the standards of guilt assessment are uncertain, the size of the legal penalties may act as a spur to litigation. Therefore, the penalties required to maintain incentives when access to insurance is unlimited may provoke too much litigation, and as a consequence, the costs of ensuring due care may decline when insurance is restricted by mandate.

On the Rationality of the Post-Announcement Drift

This paper demonstrates that a post-announcement earnings drift, which is often advanced as an example of market irrationality, can arise even if traders act rationally on their information. Specifically, we show that in the presence of share supply variations which are unrelated to information, there is a positive correlation between the unexpected component of current public signals and future price changes. Such a correlation arises from the fact that while prices reveal private information that cannot be found in public signals, non-information based trading distorts the information content of prices relative to the implications of both private and public information. Under these circumstances, markets may appear semi-strong inefficient and slow to respond to earnings announcements even though information is processed in a timely and efficient manner. Our findings correspond well with previously documented empirical evidence and suggest that the robustness of earnings-based “anomalies” may be rational outcomes of varying uncertain share supply.

Corporate Governance and Earnings Management in the Pre– and Post–Sarbanes-Oxley Act Regimes Evidence from Implicated Option Backdating Firms

Backdating stock options, a practice that retroactively adjusts stock option grant dates to lower the exercise price, has raised governance, legal, accounting, tax, and auditing concerns. The practice of backdating options generally is believed to be a result of both ineffective corporate governance and management opportunism. Both of these factors have been linked to a higher level of discretionary accruals adjustments. This study examines the accruals-based earnings management patterns for a group of firms that were implicated by the Securities and Exchange Commission (SEC) for backdating stock options with a matched control group of nonimplicated firms for a time period surrounding the enactment of the Sarbanes-Oxley Act (SOX) of 2002. Both the univariate and multivariate analyses show that in the pre-SOX years, the sample of implicated firms managed abnormal accruals at a significantly greater level than the matched group of nonimplicated firms. The differential pattern of accruals management across these two groups becomes insignificant in the post-SOX period. Our result also suggests that the effect of SOX on mitigating the level of accruals management is substantially greater for the implicated companies than for the nonimplicated companies. The difference in the effect of SOX on the two groups of firms persists even after controlling for the differences in their governance and internal control effectiveness. We, therefore, suggest that SOX had effects on management’s reporting choices beyond those resulting from improvements in governance and internal control over financial reporting.

Limits to Voluntary Disclosure in Efficient Markets

In competitive markets, prices offered by investors play a dual role: they must induce the firm to make truthful disclosures about its expected cash flows and they must also be efficient, i.e., equal the expected future cash flows to the buyer conditional on the disclosed information. We show that these requirements may exert opposing influences resulting in equilibrium disclosures being partial; that is, they might cause firms to reveal some, but not all, of the valuation relevant information possessed by the firm. We then characterize the maximal level of information that can be elicited through efficient prices. We apply our analysis to the study of voluntary disclosures in the context of equity offerings, leases and sale of tax-loss carry-forwards and compare these to the level of currently mandated disclosures under GAAP.

The Value of Correlated Signals in Agencies

We analyze optimal correlation levels in information technologies when multiple signals are available as contracting mechanisms within the principal-agent paradigm. We identify sufficient conditions ensuring that uniformly lower-correlation functions (in action levels) are preferred, as well as (mutually disjoint) sufficient conditions for a higher-correlation function to be preferred. We also show that if correlation levels are invariant in the agents action choice, the preference is for negative correlation, but not perfectly negative correlation. We generalize techniques originally used for proving Blackwells theorem and show that our results extend to the decision context as well.

Countering Opportunism in Structuring and Valuing Transactions: The Case of Securitizations

During the financial crisis of 2008, the debate on the accounting professions role has focused mostly on mark-to-market accounting rules. We argue that other weaknesses in the accounting rules as applied to securitizations predate the current crisis. The accrual system that underlies all accounting allows revenues to be recognized in advance of cash flows but requires reserves on receivables. Similarly, the fair value system books unrealized gains, but fails to recognize reserves explicitly. In illiquid markets, fair value estimates have to be based on internal models. Compensation arrangements that are closely tied to these estimates create a perfect setting for managers to increase their compensation by inflating asset values. We propose a method of recognizing fair value reserves that builds on existing practice. Such reserve accounting, an application of conservatism, will mitigate procyclical swings in fair value estimates and reduce incentives for overly optimistic asset valuations. Our analysis is conceptual in that we describe issues and their potential causes, and we propose a solution. We neither build a formal model nor perform empirical tests. Our article is exploratory and is intended to generate debate that could lead to significant improvements in accounting standards and provide a basis for rigorous empirical tests.

Normal Turbulence or Perfect Storm? Disparity in Fair Value Estimates

The use of fair value measures in financial statements embeds managerial assumptions about the future into the reporting process. This is particularly so for Level 3 measurements that are developed from financial models of future cash flows. This article documents a case where a firm bought a majority stake of 52% in November 2009 of a subsidiary where it already had an equity holding of 29% (i.e., acquired an additional 23%). The proportionate fair value implied for the 48% noncontrolling stake as well as the 29% prior equity holding in the acquired firm using Level 3 methodology was roughly triple the amount reported as Level 1 measures for these very same holdings. The discrepancy in valuation boosted the bargain gain at acquisition wiping out the retained earnings deficit of the parent firm. In addition, the acquirer reported a US

Disclosure of pension asset allocation and expected rate of return management

200 million (40%) impairment of the subsidiaries’ primary asset (housing stock) in November 2009 whereas the subsidiary reported the unimpaired value in its year-end financial statements in December 2009. While we agree that Level 3 valuations potentially provide useful information to shareholders, they can fulfill this role only if the disclosures can be effectively audited. Our primary motivation in writing this article is to show that fair value disclosures are not being audited sufficiently rigorously in practice and to make some suggestions on how these rules may be improved.

Intended or Unintended Consequences of Business Acquisitions: The Case of Financial Services Institutions

Purpose The purpose of this paper is to examine whether the expected rate of return (ERR) management is related to disclosure of pension asset allocation. FAS 132R(1), which requires firms to disaggregate the detailed categories of pension asset allocation, provides a natural experiment setting for investigating the effect of enhanced transparency on firm behavior. Design/methodology/approach The authors focus on the variation of voluntary disclosure and its effect on ERR management under the two different reporting regimes. The authors measure the variation of voluntary disclosure of the pension asset allocations in the pre-period of FAS 132R(1), by using the self-constructed disclosure score. Findings First, firms create flexibility in their choice of ERR through opaque disclosure of pension asset allocation. Next, firms with poor disclosures are more likely to adjust ERR downward when accounting standards require greater transparency, implying that, for firms with poor disclosures, mandated transparency in pension asset allocation plays a vital role in reducing the ERR management. Research limitations/implications The authors directly illustrate the impact of FAS 132R(1) on ERR management. The authors find that the impact of mandated transparency is not uniform across firms. Next, this study highlights the importance of disclosure in restricting managers’ earnings management motivation. Originality/value The authors hand collect the asset allocations under pre-FAS 132R(1) period from the 10-K pension footnotes for all S&P 500 firms, which allows the authors to identify the disclosure variation amongst the firms. Based on the variation of disclosure, the authors construct the ordinal measure of disclosure scores on which the testing indicator variables are built.

Audit quality within adverse selection markets

ASC 805 gives the management of an acquiring firm flexibility in valuation and the possibility of recognizing day one bargain purchase gains (BPG). BPG acquisitions occurred frequently in the financial services industry during the crisis of 2008 and some of these acquisitions were assisted by the FDIC which provided partially indemnification against future losses. The combination of these two events set up a unique situation to study fair value accounting. By using the data from all bank transactions, we are able to highlight the fundamental tension between relevance and reliability in fair value accounting. Our results show that in FDIC assisted transactions, fair values reflected the underlying economic transactions more accurately whereas non-FDIC assisted BPG transactions suggest that management was able to use inflated fair values to present an overoptimistic picture of the acquisition.