A Study Of The Quality And Economic Consequences Of The Accounting Information

Since January 1st 2007, the New Chinese Accounting Standards (NCAS) are implemented in all listed companies in China. Although the NCAS are converged towards the international financial reporting standards (IFRS) which pledged to be a set of high quality, single and global accounting standards, prior studies show that the high quality of standards do not always result in the high accounting quality of the listed companies. Besides, the prudential but systematic introduction of fair value in the NCAS has been regarded as a shinning spot on this accounting reform and triggers up a set of comments since its implementation. Supporters allege that fair value accounting will make the financial statements more relevant and therefore be more decision-usefulness, while the opponents worry about that the fair value accounting will increase the volatility of the accounts, leave too much room for management manipulation, make comparisons from quarter to quarter difficult and the market will overreact on fair value information and so on. Such questions deserve an empirical research and findings would be significant to further accounting reform in China.My dissertation is composed of three questions.The first one is whether the firms in the post period of application of the NCAS have a higher accounting quality than firms in the pre-period of the adoption. In particular, I investigate whether accounting amounts of firms that apply NCAS in 2007 exhibit less earnings management, more timely loss recognition, and higher value relevance than accounting amounts of firms in 2006 when they were in the pre-period of NCAS. The accounting amounts that I compare result from the interaction of features of the financial reporting system, which include accounting standards, their interpretation, enforcement, and litigation. Because my interest is in the quality of the accounting amounts that result from the financial reporting system, I make no attempt to determine the relative contribution of each of its features. I refer to the combined effect of the features of the financial reporting system as the effect of application of NCAS.The second is whether the fair value amounts are individually value-relevant. In particular, I examine whether the gains and losses on the fair value of financial assets and liabilities have incremental information contents and the factors that affect the relevance of fair value under the NCAS circumstance. To test the former, two stock price models'adjusted R-squared are compared with each other. One takes net income per-share and net assets per-share as independent variables, the other decomposes net income per-share into gains and losses on the fair value per-share and non-fair value gains and losses per-share, decomposes net assets per-share as net assets per share of the fair value of financial assets and liabilities and non-fair value net assets per-share as independent variables. To test the latter, the metric is t value. Potential factors induced from the prior literatures are test firstly in the stock price and fair value model; those with significant t value are regarded as factors affecting fair value relevance and then are multiplied by fair value gains and losses into cross items and injected into the stock price and fair value model. I base the above models on the industry fixed effect regressions.The last question I address is whether the stock returns overreact with fair value gains and losses announcement. An event study methodology is used in this section. To test this, I design five-day event windows. Estimated windows are the periods between two adjacent event windows. Raw stock returns and fair value gains and losses are regressed both on the estimation window and corresponding event window. The overreaction metric is the sum of the coefficient of fair value on the stock returns in the estimation window and in the event window.The dissertation is arranged as follows.First, it begins with an introduction. Second, analytical framework of corresponding issues is explored at chapter one. Third, the institutional sectors in China which might affect the empirical research are analyzed in chapter two. Fourth, the empirical research of accounting quality of the implementation of NCAS, the fair value and relevance, the fair value and overreaction of stock returns are carried out individually in the following chapters. Finally, Chapter six offers summary and concluding remarks.The introduction outlines the significances, the background, the methodology, thought of the issues to be discussed, the definitions of related concepts, the main contributions and the future directions needed to be studied.The Text Part is made up of six chapters.Chapter 1: First, I define the accounting quality, the relevance and reliability of accounting information, fair value and its relevance, overreaction of stock returns. Then, based on the information economics and game theory, I describe the association of accounting standards with accounting quality on the perspective of rational decision making and cost of contract. Furthermore, a framework to identify the necessary and sufficient conditions for fair value to be relevant is developed and simulated. Finally, a mechanism illustrating the overreaction implication of fair value by the magnifying channel'feeding back response is explored under a simplified financial system including households, pension funds and financial intermediaries.Chapter 2: It summarizes the characteristics of Chinese capital market, the reform of Chinese accounting standards and the corporate governance of the listed companies in China in order to figure out the institutional environment that the empirical accounting research in this paper lies in. The characteristics of"emerging and transitional"capital market, the important intervention of the government, the weak legal protection of the property wealth, the existence of the non-tradable shares, the high portion of noisy investors and the weak corporate governance in the listed companies are concluded with this chapter.Chapter 3: It starts with a literature review of the accounting quality of adoption of accounting standards. Then, a research design follows based on the review and characteristics of Chinese institutional environment drawn at chapter two. A set of comprehensive metrics of accounting quality is developed on the design. In order to get a sound empirical result, the effects of non-financial reporting system are controlled by setting a set of controlled variables, and industry fixed-effect regression is run for very metric. Empirical results are analyzed in depth at the end of the chapter.Chapter 4: It is a chapter of empirical research of fair value and its value relevance. The structure of this chapter is much like as chapter three except the issue.Chapter 5: It is a chapter of empirical research of fair value and overreaction of the stock returns. The structure of this chapter is much like as chapter three except the issue. An event study is designed and carried out industry by industry in this chapter.Chapter 6: It is a concluding section. Main findings are again summarized.The implication of the findings to standards setters, to supervisors, to investors and to accounting education is then carefully discussed. Expected direction for further study comes with the discussion of the limitations of this dissertation.Main findings in this dissertation are as follows.In controlling a serial of non-financial reporting system factors that affect accounting quality, I find that firms applying NCAS in the post-adoption year generally evidence less earnings management, more timely loss recognition, and more value relevance of accounting amounts than do firms in the pre-adoption year. In detail, according to earnings management, the variability of the change in net income scaled by total assets, the mean ratio of the variability of the change in net income to the variability of the change in operating cash flows, and the negative correlation between accruals and cash flows of the firms'evidence supports significantly less earnings smoothing of the firms at the post-adoption year than that at the pre-adoption year, but small positive net income which indicates earnings management towards a special purpose does not show lower frequency at the post-adoption year than those at pre-adoption year. Compared the earnings management in this dissertation with those in Barth et al.(2008), the listed companies'earnings management level in China is higher than the world average level regardless of the adoption of NCAS. Regarding the value relevance metrics, both the earnings per-share and net asset per-share show much higher explanatory powers to the stock price in 2007 than that in 2006, also both"good news"and"bad news"of returns explained more on the change of earnings in 2007 than in 2006. Compared with the similar research did by Barth el al.(2008), the improvements in price relevance in China after the adoption of NCAS is significant than that of the average level in the world.Summarily, more timely loss recognition and more value relevance account for most of the improvement of accounting quality in china after the adoption of NCAS. My research results also indicate that the balance sheet perspective is enhanced after NCAS's implementation. Although I cannot be sure that my findings are attributable to the change in the financial reporting system rather than to changes in firms'incentives and the economic environment, I include research design features to mitigate the effects of both.As for the fair value and value relevance, I find evidence that fair value net assets and fair value gains and losses have additional explanatory powers on the stock price. Comparing with the stock pricing theory under efficient market hypothesis, the price relevance of fair value lies mainly in the association between fair value gains and losses and stock price rather than that between fair value net assets and stock price.Empirical results also provide evidence that the market size of the firm and information transparency level has indirect influences on the value relevance of fair value. Specifically, the market size of firms affects positively on the value relevance of fair value, while information transparency level affects negatively on the value relevance of fair value. However, the age of the listed, whether firm is cross listed or not, the number of stock exchanges, local government intervening degree, financial condition, auditing quality, non-tradable ratio of the stocks and the growth of the fair value net assets don't exert significantly effects on the fair value relevance.In consistence with an analytical exploration, I find evidence that the association between fair value gains and losses and overreaction of the stock returns only exist in the financial and real estate sectors. Fair value is not the shock to overreaction, but it provides a channel to generate amplified responses.I contribute to the cooreponding literatures in the following ways.First, based on the perspective of fair value, I undertake a relatively comprehensive and thorough exploration in the improvement of accouting quality and its economic consequence after the adoption of NCAS. By using an array of quality metrics drawn from a common time period and use a common set of control variables, I provide timely and forcible evidence to support the improvement of accounting quality of implementation of the NCAS. In contrast, prior research typically focuses on individual metric such as earnings management, conservatism and value relevance. Findings from prior research comparing the quality of accounting amounts based on applying new accounting standards are mixed, which could be attributable to using different metrics, drawing data from different time periods, and using different control variables. Second, I not only demonstrate the sufficient and necessary condition of fair value to the value relevance but also bring forward a feedback mechanism of fair value's effects on the price of the property and that of the bonds. Two points of view are developed, one is that the relationship between fair value and value relevance is not natural, even when fair value is reliable. The other is the view of"feedback channel", that is, it is the fair value that provides a feedback channel which can materialize the endogenously induced increase in the property price and that of the bonds and eventually the price reach a point which is much higher than that under the historical regime.Third, I modify some of the relative methodologies for the corresponding accounting issues. In the case of the investigation on the factors impacting the relevance of fair value, I explore cross-multiple items to test the ways by which these factors affects the relevance. In order to investigate the correlation between fair value and overreaction of the stock returns, I expand the methodology event study by observing whether the sum of the coefficients in the event windows and in the estimation windows is equal to one as the benchmark of the overreaction. This is an advantage over the prior event study in which only the event windows careed about.For the contribution to the literatures of fair value accounting research, it is the first time that the"fair value gains and losses"item on the NCAS-based balance sheet becomes a subject explored in Chinese accounting literature.More importantly, I contribute to the literature of financial bubbles, it is the first time that not only an analytical framework but also the empirical evidence that the implication of an accounting regime, that is, the fair value measurement on the financial bubbles is reached on the Chinese underlying literatures.