我国上市公司财务危机预测分析

本文以上市公司被ST(特别处理)作为企业陷入财务危机的标志,运用Fisher多元判别分析、logistic回归分析和神经网络方法构建判别分析模型进行实证分析。比较其预测结果发现,神经网络方法的预测准确率明显优于Fisher多元判别分析和Logistic回归分析,因此,神经网络方法更适合于企业财务危机预测。     

基于Fisher判别法的P2P网络借贷平台信誉等级评价模型

P2P网络借贷是一种新兴的互联网金融,平台的信誉是影响投资者选择平台的主要影响因素之一。选取成交积分、人气积分、营收积分、分散积分、杠杆积分、透明度、品牌、流动性、收益积分9个影响因素作为判别平台等级预测的评价指标,同时对这9个评价指标做主成分分析,提取出3个主要成分,用Fisher判别法进行训练预测,建立了基于主成分分析的网络借贷平台信誉等级预测的Fisher判别模型。通过对“网贷之家”公布的37组平台数据作为训练样本数据集进行模型的训练,12组数据作为该预测模型的测试数据,进行网络借贷平台信誉等级的预测,同时通过其他预测模型预测结果的对比,验证了Fisher判别法在网络借贷平台信誉等级的预测中具有较低的误判率,其误判率仅为1/12。     

Z-score模型最优分割点的确定方法比较——基于违约风险预测能力的分析

Z-score模型在对企业进行财务困境和违约风险判别方面具有重要的应用价值,最优分割点的确定方法对于提高模型的违约风险判别能力至关重要.本文以医药生物行业上市公司为样本,运用Fisher逐步判别法从15类财务比率中筛选出判别能力较强的7个指标构建了Z-score模型,并尝试采用加权平均法和考虑先验概率及误判成本的ZETAc模型法分别确定最优分割点.研究发现,ZETAc模型法预测企业违约风险的能力明显优于加权平均法.     

基于Bow-tie模型的模糊贝叶斯网络在风险评估中的应用

本文以Bow-tie模型为基础,借鉴工业部门风险分析方法,建立基于Bow-tie模型的模糊贝叶斯网络风险评估框架,实现从静态层次分析向动态网络推理转变、从风险评估向风险预测深化、从风险发生前评估向风险发生后分析发展,为内部审计职能拓展与价值深化提供借鉴。     

行业与区域的企业财务危机预警模型比较研究

本文运用多元判别模型、Logit模型、主成分模型,对不同行业和地区的企业进行财务危机预警研究,进而分析判别准确率的差异.同时,对预警模型的指标选择和不同类型危机的预警判别进行了比较分析.本文研究生成的预警模型可以提供给商业银行进行风险度量使用.     

基于多元判别方法的商业银行信用风险度量研究

本文以我国机械行业上市公司为研究样本,运用统计分析中的多元判别方法建立了预测企业信用风险程度的判别模型,研究结果表明模型具有较高的判别准确率。该模型适用于我国商业银行对机械行业企业的信贷分析,对商业银行建立和优化自身的信用风险度量模型也具有一定的参考价值。     

我国上市公司财务危机预警模型的实证研究

本文以我国上市公司指标为研究对象,以25家因财务状况进入特别处理的公司和25家财务正常的上市公司为样本,选取了6个变量指标,通过单变量模型,一元判定模型,Fisher判定模型和logistic模型四种分析方法对公司的财务状况进行实证分析。单边变量立面分析模型和一元判定模型运用单变量对财务危机进行简单预测,误差大,预测能力差,而Fisher模型和logistic模型从多变量进行预测,预测能力明显比单变量强,其中logistic模型预测能力最强。     

基于贝叶斯网络的上市证券公司风险预警模型研究

上市证券公司的风险预警模型能够为政府监管、证券公司稳健发展以及投资者研判提供依据。以上市证券公司风险管理指标体系为基础,利用贝叶斯网络方法以及支持向量机、随机森林和多项Logit模型分别建立风险预警模型进行比较,并在实证中针对上市证券公司的不平衡数据特征,用 SMOTE抽样对数据进行预处理。最终实证表明：从平均准确率和标准差两个角度比较,SOMTE抽样增加了贝叶斯网络的预测效果,机器学习方法要优于多项Logit模型,贝叶斯网络方法效果最佳。     

基于集成算法的信用债违约预测模型及其影响因素研究

本文选取2014—2021年的信用债作为研究对象，以单一算法（逻辑回归、高斯朴素贝叶斯、支持向量机和伯努利朴素贝叶斯）分别构建Bagging和Stacking集成学习模型，并将其与单一算法以及传统方法（KMV模型）进行对比，探讨如何提高信用债违约预测精度，证明Bagging集成算法的可靠性，还通过SHAP可解释算法研究信用债违约中的重要指标。实证结果显示，四种单一机器学习模型在预测准确率上优于传统KMV模型；进一步对机器学习模型进行集成，发现同质集成算法在提高预测性能方面不如异质集成算法，且Bagging异质集成算法的预测准确率优于Stacking异质集成算法。将性能最好的Bagging异质集成算法与SHAP可解释算法结合，得到对识别信用债违约具有重要价值的9个指标，分别是违约前债项评级、违约前主体评级、交易所、净资产收益率、债券类型、企业性质、财务费用、总资产增长率以及第一大股东持股比例，关键指标的识别对信用债违约预测具有指导意义。     

上市公司财务危机预警的实证研究

我国企业必须建立财务预警系统,以此帮助企业地预测、防范财务危机。本文选取了我国沪两市上市公司2008-2009年的80家ST公司和845家非ST公司作为研究样本,以2006--2007年的财务指标作为模型研究的变量,对上市公司财务危机进行预警研究。本文选择了8个财务指标作为建立模型的变量,然后采用统计中的贝叶斯判别分析方法,建立了上市公司财务危机的判别分析识别模型,模型预测效果良好,具有实际应用价值。     

Debt and Taxes and Uncertainty

With a graduated personal tax schedule, Miller showed that there could be an equilibrium debt supply for the corporate sector as a whole. In the presence of uncertainty there is also a unique debt/equity ratio for each individual firm, and this ratio is related to the firm's operational risk characteristics. However, if firms merge and spin off in response to tax incentives, the identity of firms is ambiguous and only the corporate sector is a meaningful construct. These arguments are developed in both discrete and continuous models that employ extensions of the arbitrage-free pricing theory.     

教育与资本怎样的前景怎样的现实

来自政府的声音: "动员各种社会资源,发展教育" 此次论坛上财政部长助理张少春的发言是最受关注的发言之一.他的演讲传达出的信息有这样几方面:第一,谈到现实,他认为全社会对教育的巨大需求与我国公共投入不足已成为我国教育事业发展的突出矛盾.近年来,以公共部门投入为主,多渠道筹措教育经费的教育投入机制成为解决这一问题的重要途径,也使得公共部门与私营部门在教育领域的合作日益密切.第二,谈到前景,他认为,1.中国经济持续、快速的增长对高技能人才产生巨大的需求,这为私营部门的参与提供了广阔空间.2.随着公共财政职能的不断完善,财政资金将在各项教育事业中重新进行分配,进一步优化财政支出结构,提高资金的使用效率.在"十一五"期间,我国将把公共支出的重点转移到农村.我们将逐步把全体农村适龄儿童的义务教育全部纳入公共财政体制.     

East and west and north and south : A preliminary model

The Brandt report and other proposals for a new North-South world order continue to be the focus of lively debate, yet it is often argued that little concrete has emerged from their recommendations. A major reason for this, the author argues, is that the East-West conflict component is not sufficiently taken into consideration-East-West tensions are a severe constraint on and a determining factor of North-South relations. Using influence diagrams, the author develops a flexible framework for discussion and assessment of N/S-E/W relations.     

Trust and cross-border mergers and acquisitions

A number of studies suggest that social trust matters for investment. Using different measures of trust from World Values Survey, we show that countries where people display higher levels of trust engage in more cross-border M&A activities. When they do, these acquirers pay lower premiums. To the extent that these acquirers also tend to engage in larger acquisitions as well, our findings suggest that a larger selection pool of potential targets and higher value targets enable these acquirers to negotiate for lower premiums. We do not find evidence of the significant effect for target country trust levels. Hence, trust may benefit those acquiring firms in cross-border transactions.     

Tournament-based incentives and mergers and acquisitions

This research examines the relation between tournament-based incentives, which are proxied by the difference between a firm's CEO pay and the median pay of the senior managers, and mergers and acquisitions (M&As). We find that tournament-based incentives are positively related to firm acquisitiveness and acquiring firms' stock and operating performance. Further analysis indicates that positive acquisition performance increases the likelihood of the CEO being promoted from inside the acquiring firm. Our evidence is consistent with the view that tournament-based incentives motivate acquiring firms' managers to make greater efforts and take more risk that result in superior acquisition performance.     

Corporate Innovations and Mergers and Acquisitions

Using a large and unique patent‐merger data set over the period 1984 to 2006, we show that companies with large patent portfolios and low R&D expenses are acquirers, while companies with high R&D expenses and slow growth in patent output are targets. Further, technological overlap between firm pairs has a positive effect on transaction incidence, and this effect is reduced for firm pairs that overlap in product markets. We also show that acquirers with prior technological linkage to their target firms produce more patents afterwards. We conclude that synergies obtained from combining innovation capabilities are important drivers of acquisitions.     

Director gender and mergers and acquisitions

Does director gender influence CEO empire building? Does it affect the bid premium paid for target firms? Less overconfident female directors less overestimate merger gains. As a result, firms with female directors are less likely to make acquisitions and if they do, pay lower bid premia. Using acquisition bids by S&P 1500 companies during 1997–2009 we find that each additional female director is associated with 7.6% fewer bids, and each additional female director on a bidder board reduces the bid premium paid by 15.4%. Our findings support the notion that female directors help create shareholder value through their influence on acquisition decisions. We also discuss other possible interpretations of our findings.