可轉債近年來已成為我國公司進行籌資活動時的主要工具之一，但可轉債的宣告效果多被市場視為不利的消息，發行後之長期績效也呈現持續惡化的趨勢，除了市場反應不足外，發行公司進行盈餘管理亦為重要的影響因素之一；公司於可轉債發行前利用盈餘管理的方式提升績效，將造成發行後產生盈餘反轉的效果，因而使得長期營運績效與股價表現不佳。而操弄盈餘的行為在較嚴格的監督機制與完善的內部治理制度下，能夠因為外部約束的力量以及公司內部的自律機制而有所減緩，故進一步利用公司治理變數檢視對盈餘管理的影響，並探討國內可轉債與海外可轉債兩者間因為投資人區隔、發行市場差異的影響，對於其盈餘管理以及發行後長期績效的影響。 本研究之主要之實證結果如下: 1. 可轉債發行之宣告效果皆為負向，但海外可轉債之異常報酬率的惡化程度不如國內可轉債來的嚴重，其發行之訊號效果優於國內可轉債；長期而言，不論國內可轉債或海外可轉債發行後之長期股價報酬皆為顯著的負值，於發行後三年內皆呈現持續惡化的情形，與先前相關文獻之結果一致。 2. 國內可轉債與海外可轉債發行前夕公司皆有刻意進行盈餘管理的行為，並於發行後發生盈餘反轉使得公司營運績效下滑；利用多元迴歸分析觀察盈餘管理對長期績效的影響，發現公司於可轉債發行前進行盈餘管理對其發行後之長期績效有顯著的不良影響，而海外可轉債發行公司由於受到外國規章監督，其進行盈餘管理將受到更為嚴重的處罰，導致股價下跌的幅度較大。 3. 進一步以多元迴歸分析檢視公司治理對於盈餘管理的影響，發現當發行公司的治理制度越佳，代理問題越小時，較不會在可轉債發行前進行盈餘管理來欺瞞投資人；而可轉債發行地點亦為影響盈餘管理程度的另一項因素，當可轉債為海外發行時，發行公司進行盈餘管理的幅度顯著小於國內發行的公司。

The thesis processes the auditor’s role and whether it can constrain possible earnings management in smaller Swedish private corporates. The removed auditing obligation that was abolished in Sweden 2010 with the reason to reduce the administrative burden faced by smaller corporates, contributed to an important discussion regarding the auditor’s actual importance. In addition, the corporate tax in Sweden was significantly reduced from 26.3 percent to 22 percent, which gave Swedish corporates incentives to reduce their earnings before the tax reduction, and therefore pay less taxes. This phenomenon is used in the thesis as an excellent opportunity to measure earnings management since there were significantly strong incentives for corporates to manipulate their earnings at that time. The thesis aims to increase understanding among these corporates accounting, and whether the auditor can strengthen the accounting- and audit quality, and therefore constrain earnings management. Through the use of a quantitative method where statistical tests have been performed on underlying data gathered from the corporates annual reports, the thesis research question has been answered. The formulation of the research questions is “Does the degree of earnings management differ between audited and unaudited corporates?”, which has been investigated by measuring the degree of earnings management through unexpected accruals, but also through SG&A cost stickiness. Furthermore, the thesis aims to provide the research area with arguments regarding the appropriateness of the audit exemption. The result obtained in the thesis indicates that audited corporates have less negative unexpected accruals, and hence a minor degree of earnings management in comparison with unaudited corporates. However, no significance is shown in the result, at a five percent significance level, which means that it’s not possible to say with certainty that there’s a difference between these two groups in terms of negative unexpected accruals. Instead, there is evidence that there’s a significant difference between audited and unaudited corporates in terms of cost stickiness, where unaudited corporates showed more cost stickiness and thus also a higher degree of earnings management. With an additional test the thesis also demonstrates that there are corporates that don’t follow the Swedish laws regarding audit exemption. This is when it’s discovered that there are corporates that aren’t covered by the audit exemption, and thus don’t meet the required limits, but still don’t provide an auditor. These corporates also prove to have a higher degree of earnings management since the thesis received a significant result in terms of negative unexpected accruals. To sum up, the thesis highlights that the auditor’s role is of great importance in several aspects, and for this reason there are motives for legislators to review the audit exemption in Sweden.

民國八十七年亞洲金融危機引發本土性金融風暴後，集團企業交叉持股在掌握經營權、炒作股價及企業間利益輸送所造成之負面經濟影響，再次成為議題的焦點。在台灣證券管理交易法規中，企業交叉持股行為僅能由母公司代子公司公告窺知一二。國內文獻中對於交叉持股之探討亦偏重法律制度及盈餘管理層面，對於透過交叉持股之資金操縱行為則較少著墨。 有鑒於交叉持股「能載舟亦能覆舟」之經濟重要性，及現行公告實務未能窺知交叉持股行為全貌，本研究嘗試以投資餘額變動頻率異常作為潛在的交叉持股行為代理變數，將民國八十三年至八十八年第三季之台灣上市公司按投資餘額變動之頻繁程度區分為高變動群與低變動群樣本，進行其財務性操縱行為的探討。本研究主要的實證議題有二：（1）投資餘額變動較頻繁之公司其盈餘管理行為是否異於投資餘額變動較不頻繁之公司？及（2）投資餘額變動較頻繁之公司其資金操縱行為是否異於投資餘額變動較不頻繁之公司？ 實證結果顯示，在盈餘管理行為上，高投資變動公司有顯著利用裁決性應計項目操縱盈餘的現象，且此現象在曾發生財務危機的高投資變動公司更為顯著。高投資變動公司在多頭市場時，投資收益顯著較低投資變動公司為大；空頭市場時，兩者在真實盈餘操縱變數上並無顯著差異。在資金操縱方面，高投資變動公司質押成數變動率顯著大於低投資變動公司，且曾發生財務危機的高投資變動公司變動程度更為明顯。此外，高投資變動公司債務活動在多頭時期較低投資變動公司活絡，顯示債務活動活躍是高投資變動公司重要特性之一。基於現行法規實務與資料來源的限制不易直接偵測具經濟實質之負面交叉持股行為，本研究之綜合實證發現高投資變動公司與真實負面交叉持股行為有相當程度的關聯性，是故投資餘額變動頻率異常似乎是研究負面交叉持股公司極為重要的參考指標之一。

本論文以公司規模大小、公司成長率、盈餘變異程度、盈餘持續率、負債比率、系統風險、以及產業別為控制變數，探討管理當局持股比率與管理當局自願性（強制性）盈餘預測準確度、盈餘管理程度之關係。並進一步探討管理當局持股比率與七個控制變數之交互作用對管理當局自願性（強制性）盈餘預測準礁度及盈餘管理程度之影響。 　　本實證研究結果發現： 　　1、自願性盈餘預測方面： 　　(1)管理當局持股比率越高且盈餘變異程度越大之公司，盈餘預測誤差越高，盈餘預測準確度越低。 　　(2)管理當局持股比率越高且負債比率越高之公司，盈餘預測誤差越高，盈餘預測準確度越低。 　　(3)產業別會影響其預測準確度，而產業中以鋼鐵業之盈餘預測準確度，顯著較高。 　　(4)公司成長率越高、盈餘持續率越高，其盈餘管理程度越高。 　　(5)產業中以電子業有顯著較高之盈餘管理程度。 　　2、強制性盈餘預測力面： 　　(1)管理當局持股比率與盈餘預測準確度成正相關。 　　(2)公司規模與盈餘預測準確度成負相關。 　　(3)盈餘持續率與盈餘預測準確度成負相關。 　　(4)產業別確實與強制性盈餘預測準確度有關，其中以電子業之盈餘預測準確度顯著較低。 　　(5)管理當局持股比率越高之紡織業其盈餘預測準確度顯著較低。 　　(6)強制性盈餘預測並無顯著的盈餘管理情況產生。 　　3、綜合結論： 　　(1)自願性之盈餘預測準確度高於強制性之盈餘預測準確度。 　　(2)自願性之盈餘管理程度高於強制性之盈餘管理程度。 / This research hypothesizes that the level of managerial ownership that controlling for earnings growth、earnings variability、earnings persistence、company risk、 debt、industry、and size has effect on both the magnitude of forecast precise of voluntary(compelling) forecast and the magnitude of discretionary accounting accrual adjustment. 　　In addition，this study examines that there are interaction of ownership effects on both the magnitude of forecast precise of voluntary (compelling) forecast and the magnitude of discretionary accounting accrual adjustment. 　　The empirical results show as follow: 　　1、Voluntary forecast aspect: 　　(1) Managerial ownership is negatively associated with the magnitude of forecast precise. 　　(2) Managerial ownership of is positively associated with the magnitude of discretionary accounting accrual adjustment. 　　2、Compelling forecast aspect: 　　(1) Managerial ownership is positively associated with the magnitude of forecast precise. 　　(2) Managerial ownership is not associated with the magnitude of discretionary accounting accrual adjustment. 　　3、Conclusion explication: 　　(1) The magnitude of forecast precise of voluntary forecast is more than that of compelling forecast. 　　(2) The magnitude of discretionary accounting accrual adjustment of voluntary forecast is more than that of compelling forecast. 　　(3) Industry variable indeed affects both the magnitude of forecast precise and the magnitude of discretionary accounting accrual adjustment.

本研究分成兩部分，研究對象為中國大陸之上市公司，第一部分以2004至2009年之財務資料，探討2008年中國大陸企業所得稅法之實施對中國大陸上市公司平均有效稅率之影響；第二部分以1998至2009年之財務資料，探討中國大陸上市公司是否為了因應企業所得稅法實施對有效稅率的影響，而從事租稅誘因盈餘管理之行為。 第一部分之實證結果發現，中國大陸上市公司之平均有效稅率在2008年企業所得稅法實施後有顯著的下降，表示中國大陸上市公司之有效稅率隨著名目稅率降低而下降。外資持股比例超過20%之上市公司（定義為外資企業），其2008年後之有效稅率上升，且內外資企業之其平均有效稅率之差異有縮小之趨勢。屬於高科技產業之上市公司在2008年企業所得稅法實施後，其有效稅率並未下降；但與除了農林牧漁業外之其他產業相比，高科技產業之上市公司其有效稅率較低。第二部分之實證結果發現，在2008年有效稅率下降之上市公司，其於2008年後的裁決性應計項目較大，有將盈餘遞延認列之傾向。 / This study consists of two research issues. The first part uses the financial statement and capital structure data of China listed companies from 2004 to 2009, to investigate the effects of the implementation of the Law of the People’s Republic of China on Enterprise Income Tax in 2008(hereafter 2008 Income Tax Law). And the second part which uses the financial and capital structure data of China listed companies from 1998 to 2009 investigates if China listed companies take the action of earnings management in response to the implementation of the 2008 Income Tax Law. The empirical findings of the first part show that the average ETRs of China listed companies declined after the implementation of the 2008 Income Tax Law. The listed companies whose proportion of foreign ownership is above 20% (defined as foreign enterprises) increased ETRs after 2008 and the differences in ETRs between domestic enterprises and foreign enterprises are reduced gradually. Further, the average ETRs of China listed companies of information technology industry do not appear to be decreased after the implementation of the 2008 Income Tax Law. However, besides the agriculture, forestry, herding, fishery industries, listed companies of information technology industries remain to have lower average ETRs after implementation of the 2008 Income Tax Law. The empirical findings of the second part show that China listed companies whose average ETRs decreased in 2008 appeared to have greater discretionary accruals after 2008, suggesting those companies shifting earnings to 2008.

The objective of the International Financial Reporting Standards (IFRS) is to provide useful information to the users of financial statements to assist in making economic decisions. To be useful, information has to be relevant and reliable, but the reliability of information suffers when the guidelines for the reporting of specific issues are not clear and managerial discretion arises. Write-offs are one of those accounting issues that are regularly related to earnings management. By now it is seen as common knowledge that write-offs, especially those on goodwill, do not reflect declines in asset value; rather, they are used as a device to manipulate financial reports. However, there is a striking lack of grounded theoretical research that can confirm this assessment. The aim of this dissertation is to provide valuable analytical and empirical insights on fixed asset write-offs under IFRS. In a first step, the practical implementation of IAS 36 in Europe has to be analyzed, which is best done empirically. Based on the findings from these empirical surveys, the most substantial questions remaining are subject to an in-depth analytical discussion. Since IAS 36 entails different measurement issues that have their origins in finance theory, this dissertation also aims to introduce some basic techniques from theoretical finance to accounting research. Lastly, as the analyses presented in this dissertation do not cover all open questions on fixed asset write-offs, the author hopes to encourage further research on this important topic.

Three essays are comprised in this dissertation to examine how institution and regulation frameworks affect the quality of financial reporting by banks. The empirical investigation on whether some governance mechanisms provide incentives for banks to report high quality financial information can have policy implications regarding bank regulation. Financial reporting quality is measured either by the level of earnings management or the extent of reporting conservatism. Using these two types of proxies for financial reporting quality, we examine whether reporting quality is affected by the legal protection on investors, bank supervision/regulation practices, or securities market governance mechanisms. In the first essay, we examine international differences in bank earnings management around the world. Following Leuz et al. (2003), we argue that bank earnings management is closely linked to private benefits of insiders. As a result, bank earnings management should be negatively related to institutional factors such as legal protection on investors and bank supervision policies that encourage market discipline on banks. Consistent with this prediction, we provide evidence that earnings management is less pervasive for banks in countries where investors are better protected and where supervision policies strongly encourage private-sector monitoring on banks. We also show that the legal protection mechanisms have stronger effects on curbing activities of earnings discretion, but bank supervision policies that encourage private-sector monitoring are better at limiting income smoothing activities. Our results also suggest that stringent capital requirement or strong government supervisions are less effective in reducing earnings activities of banks. In the second essay, we document that banks, especially those that are publicly traded, are conservative in their financial reporting. In particular, banks are conservative in reporting earnings changes and they incorporate more loan loss provisions when their operating cash flows decrease or when the amount of their problem loans increases. Banks also charge off more problem loans when their loan loss provisions increase. Our cross-country comparison shows that conservative financial reporting is more pronounced in countries where supervisors are empowered to take adequate actions against banks or where bank supervisory policies to encourage private-sector monitoring are more prevalent than in countries where there is less supervision or where there is less private-sector monitoring. In the third essay, we further investigate whether securities market governance explain the international differences of reporting conservatism across listing status of banks. Our results indicate that, after controlling for banking industry regulations, securities market governance has incremental effects on the reporting conservatism by public banks. The conservative reporting by public banks is stronger in countries where securities regulators are more empowered to intervene in banks for violations to securities laws. Furthermore, the stronger conservatism for public banks relative to private banks is widespread in countries with more developed bond market. The evidence suggests that public banks practice more conservative reporting than their private counterparts when debt contracting mechanisms function well.

Nuvarande redovisningsregler och principer under IFRS är till viss del öppna för subjektiva bedömningar om hur kostnader och intäkter ska redovisas. Denna subjektivitet går stundom under benämningen resultatmanipulering och innebär då att företagets resultat antingen justeras i positiv eller negativ riktning beroende på bakomliggande incitament. De subjektiva valen kan delas in i operationella åtgärder eller diskretionära periodiseringar, beroende på om de påverkar kassaflödet eller ej. Denna studie fokuserar på hur redovisningsreglerna kan användas för att uppnå resultatmanipulering och inriktas därför på de diskretionära periodiseringarna. Forskning har påvisat att möjligheten till resultatmanipulering är särskilt påtaglig i samband med att företag byter vd. Det baseras på teorierna om horisontproblemet och stålbad, att vd:n i varje given situation ämnar uppnå maximal personlig vinning. Då tidigare studier sällan undersökt den svenska marknaden avser vi uppnå kontribution om hur teorierna om resultatmanipulering är applicerbara under svenska regelverk och normer, från dess att IFRS infördes till dags dato. Via syftet att undersöka sambandet mellan ett byte av vd och resultatmanipulering har vi för avsikt att hjälpa investerare till att bättre förstå marknaden och till förbättrade investeringsbeslut. Via kvantitativ analys av årsredovisningar har vi med hjälp av den modifierade Jonesmodellen tagit fram ett mått på företags diskretionära periodiseringar. Måttet har likställts med resultatmanipulering och ställts i relation till företags vd-byten. Kopplat till horisontproblemet har nivån på resultatmanipulering jämförts mot vd-byte som sker efterföljande räkenskapsår. Teorin om stålbad har i sin tur testats genom att undersöka om resultatmanipulering nyttjas under en vd:s första verksamma år, alternativt i något av de efterföljande två räkenskapsåren. Resultaten påvisade att den svenska marknaden skiljer sig mot både bakomliggande teori och tidigare forskning. Förekomsten av horisontproblemet eller stålbad kunde inte bevisas, däremot fann studien att resultatet skrivs upp via resultatmanipulering i de två efterföljande åren efter att ett byte av vd skett. Anmärkningsvärt bevisades även att resultatminskande åtgärder utförs året innan ett rutinmässigt vd-byte inträffar. Sett till orsakssambanden bevisades därutöver att sannolikheten för att ett vd-byte kom att inträffa ökade vid användning av negativ resultatmanipulering samt när företagets lönsamhet var lägre än marknadsgenomsnittet. Studien belyser att kausalitetsproblem föreligger mellan resultatmanipulering, vd-byte och företagets lönsamhet. De exakta orsakssambanden kan studien inte förmå förklara utan vi påvisar att ytterligare forskning behöver genomföras för att klargöra det sanna orsaksförhållandet. Avslutningsvis renderar studien teoretiskt bidrag då den påvisar att teorierna om horisontproblemet och stålbad i samband med vd-byte inte stämmer in på den svenska marknaden.

This thesis comprises three studies on CEO compensation in the UK. It specifically examines the role of CEO defined-benefit pensions, compensation consultants and CEO stock options. Firstly, research on the role of executive pensions is still at a stage of infancy due to data difficulties (Sundaram and Yermack, 2007). By taking advantage of better disclosure requirements with the introduction of Directors’ Remuneration Report Regulations (DRRR) in 2002, this thesis examines the determinants and effects of CEO defined-benefit pensions. Consistent with rent extraction hypothesis (Bebchuk and Fried, 2005; Kalyta and Magnan, 2008), it finds that pensions are largely determined by CEO power over boards of directors. There is no evidence that pensions reduce the agency cost of debt as suggested by Edmans (2008) and Sundaram and Yermack (2007). Instead they increase the agency cost of equity by discouraging CEO risk-taking and reducing pay-performance relationship. Consistent with the argument in Gustman et al. (1994), Ippolito (1991) and Lazear (1990), this thesis also finds that pensions do bond a CEO to the firm she manages. Secondly, because of the lack of disclosure regarding compensation consultants used by companies, the empirical evidence is so far limited on how the practice of employing compensation consultants influences CEO pay. By taking advantage of better disclosure requirements since the publication of the DRRR (2002), this thesis examines the effect of using compensation consultants on CEO pay. Unlike Murphy and Sandino (2008), this thesis finds no evidence that firms use multiple pay consultants to justify or legitimize higher CEO pay. In light of the managerial power theory, this thesis instead finds that pay consultants are more concerned with the risk of losing business with their client firms. This latter finding explains why the use of pay consultants is associated with greater executive pay (see Armstrong, Ittner and Larcker, 2008; Cadman, Carter and Hillegeist, in press; Conyon, Peck and Sadler, 2009; Murphy and Sandino, 2008; Voulgaris, Stathopoulos and Walker, 2009). Thirdly, despite the importance of the issue, the existence of a link between the CEO stock options and earnings management is currently understudied in the UK. The UK context is appealing because of two distinctive corporate governance features that limit opportunistic earnings management. These are the absence of CEO duality in general (Cornett, Marcus, and Tehranian, 2008) and the increased outside director’s membership on boards since the publication of the Cadbury Report (1992) (Peasnell, Pope, and Young, 2000). By examining earnings management prior to stock option grant and exercise periods, this thesis adds to the study of Kuang (2008) that examines earnings management during stock option vesting periods. Overall, some evidence has been found that earnings are managed downwards prior to stock option grant periods. Consistent with the US-based studies, this thesis finds strong evidence of upward earnings management prior to a stock option exercise period. It shows that the UK’s distinctive governance features have not restrained opportunistic earnings management prior to stock option grants and exercises. In brief, this thesis provides some empirical evidence that the use of two pay components in the CEO pay package, namely, the defined-benefit pensions and stock options, do not necessarily promote CEO-shareholder interest alignment. The use of pay consultants in CEO pay-setting is also fraught with managerial influence. In support of the managerial power theory, I therefore suggest that these three factors can be abused by CEOs to extract excess compensation at the expense of shareholders. In this context, these three factors can themselves be considered as the sources of the agency cost. Future research might examine the mechanisms that can be deployed to govern the use of defined-benefit pensions, stock options and pay consultants in CEO pay design.

會計盈餘是企業向投資人傳達營運績效的指標，又投資人視公司達成分析師預測門檻與否為企業前景的重要訊號。當公司宣告策略聯盟決策時，資本市場給予正面評價，不過策略聯盟協議可能使管理當局存在機會主義與盈餘管理活動，過去文獻發現，有策略聯盟的公司，其盈餘品質較低。本文探究企業執行策略聯盟對於跨越盈餘門檻的關聯性，觀察策略聯盟事件是否為管理當局進行盈餘管理或是預期管理的工具，以迎合或擊敗分析師之盈餘預測門檻。本文實證發現策略聯盟與否及策略聯盟多寡與分析師預測門檻具有顯著正相關，而執行策略聯盟之公司從事向下引導分析師預測的機率較低，此外，實證結果亦發現，策略聯盟會降低公司管理當局採取向上調整裁決性應計數的可能性。綜上研究顯示，有策略聯盟之企業達成分析師預測門檻的機會較高，然而，其管理當局較不會選擇應計項目盈餘管理或預期管理方式來迎合或擊敗分析師之盈餘預測門檻。