会计稳健性研究

【作者】 李远鹏；

【导师】 李若山；

【作者基本信息】 复旦大学 ， 会计学， 2006， 博士

【摘要】 本文对中国上市公司的条件稳健性和非条件稳健性进行了系统研究。本文首先以沪深股市的上市公司为样本，研究了亏损公司的利润操纵行为对条件稳健性检验的影响；其次，本文对亏损公司行为解释的一些竞争性假说进行了检验；最后，本文研究了上市公司条件稳健性和非条件稳健性的时间序列特征，并发现，中国上市公司从1998年开始稳健性在提高，而这种提高主要源于非条件稳健性的提高。对于条件稳健性，本文在第三章中以1998年到2004年中国上市公司为样本，运用了盈余的偏度和波动程度、盈余-股价回报关系、盈余的持续性模型、盈余对现金流信息增量、会计应计-现金流量关系等6个代理变量研究了其存在性问题。本文发现样本整体表现出了稳健性的特征，但是当控制住亏损公司的影响后，这些稳健性的特征都消失了，这就意味着整体的稳健性是亏损公司造成的。对于亏损公司所表现出的这一特征，主要有三种解释：卖出期权观、稳健性观和利润操纵观。对于这三种解释在本文的第四章中进行了逐一的分析。首先，对于卖出期权观而言虽然它可以解释本文发现的某些现象例如盈余-股价回报关系，盈余-现金流信息增量等，但是却无法解释其它的一些发现，例如，盈余的波动程度和偏度等。其次对于稳健性观，我们认为它无法解释盈余-现金流信息增量这一实证发现，并且我们还利用了稳健性的契约解释，其结果都无法支持契约观。最后对于利润操纵观，我们认为它可以解释我们全部的经验发现。本文的第五章对中国上市公司稳健性的时间序列特征进行了研究。本文首先发现中国上市公司从1998年到2004年ROA一直在下降，而收入、主营业务利润、经营活动现金流量却都在上升，这说明了ROA的下降是会计应计造成的。本文又将应计分为经营性应计和非经营性应计，发现从1998年到2004年应计的下降主要受到了非经营应计的影响，而且至少从2001年开始非经营应计一直为负。以ROA分组研究应计下降的趋势，支持应计的下降主要是非条件稳健性上升的结果。基于盈余的波动和偏度以及盈余股价关系的证据也进一步支持了这一论点。更多还原

【Abstract】 This dissertation studies the accounting conservatism, including conditional conservatism and unconditional conservatism, in China mainland. Firstly, I find that the earnings of listed companies reflect bad news more quickly than good. Using six variables to measure conservatism, we find that its presence in the full sample results from the big bath behavior of loss companies. Secondly, I test the alternative hypothesis about big bath behavior of loss companies, including put option theory and accounting conservatism theory, and prove that the earnings management theory can explain all the facts that I has found. Finally, using data from 1994 to 2004,I test the time series properties of earnings, cash flows and accruals, and find that the financial reports become more and more unconditionally conservative in China mainland.The topic of Chapter 3 is conditional conservatism. Using the companies listing in SHSE and SZSE form 1998 to 2004, I find that there is accounting conservatism in China mainland: the earnings of listed companies reflect bad news more quickly than good. Because of the ST and delisting requirements in China’s mainland, loss companies have strong incentives to take a big bath, which leads to bad news being reflected more quickly than good news in accounting earnings. Using 6 proxy to measure conditional conservatism, including the skewness of earnings distribution, the variability of earnings, the earnings-return association, the accruals-cash flow association, the persistence of earnings, the information content of earnings relative to cash flow, I find that the conditional conservatism is gone when I control the effect of the loss companies. Witch means that the conservatism of the full samples is the result of the big bath behavior of loss companies and there is no conditional conservatism in China.There are alternative explanations for big bath behavior itself, including the abandonment option and conservatism. In Chapter 4, I attempt to provide some additional discussion of these explanations. The abandonment option theory cannot explain why the earnings volatility of loss companies is higher than that of profitable concerns., and it either cannot explain the negative skewness of accounting earnings and the negative distribution of ROE. The conservatism can’t explain the why the conservatism depends on operating performance. But the earnings management can explain all the facts that I has found. So I can draw the conclusion that conservatism in China mainland is the result of the big bath behavior of loss companies.In Chapter 5, we test the time-series properties of earnings, cash flows and accruals. Firstly, I find that ROA is declining since 1998, while the CFO, sales and major operating profits are going up from 1998. So the declining of ROA is the result of the accruals, which means that financial reporting become more conservative in China mainland. Further more, I prove that the conservatism is unconditional conservatism. Finally, Using 3 proxy to measure conditional conservatism, I find that there is no conditional conservatism in China when I control the big bath behavior of loss companies.更多还原