Data envelopment analysis in the presence of measurement error: case study from the National Database of Nursing Quality Indicators® (NDNQI®)

Data envelopment analysis (DEA) is the most commonly used approach for evaluating healthcare efficiency [B. Hollingsworth, The measurement of efficiency and productivity of health care delivery. Health Economics 17(10) (2008), pp. 1107–1128], but a long-standing concern is that DEA assumes that data are measured without error. This is quite unlikely, and DEA and other efficiency analysis techniques may yield biased efficiency estimates if it is not realized [B.J. Gajewski, R. Lee, M. Bott, U. Piamjariyakul, and R.L. Taunton, On estimating the distribution of data envelopment analysis efficiency scores: an application to nursing homes’ care planning process. Journal of Applied Statistics 36(9) (2009), pp. 933–944; J. Ruggiero, Data envelopment analysis with stochastic data. Journal of the Operational Research Society 55 (2004), pp. 1008–1012]. We propose to address measurement error systematically using a Bayesian method (Bayesian DEA). We will apply Bayesian DEA to data from the National Database of Nursing Quality Indicators^® to estimate nursing units’ efficiency. Several external reliability studies inform the posterior distribution of the measurement error on the DEA variables. We will discuss the case of generalizing the approach to situations where an external reliability study is not feasible.     

Efficiency measures of the Chinese commercial banking system using an additive two-stage DEA

Measuring and improving the efficiency of the Chinese commercial banking system has recently attracted increasing interest. Few studies, however, have adopted the two-stage network DEA to explore this issue in the Chinese context. Because the entire operational process of the banking system could be divided into two sub-processes (deposit producing and profit earning), the evaluation of the sub-process efficiencies could be used to assist in identifying the sources of the inefficiency of the entire banking system. In this study, we utilize the network DEA approach to disaggregate, evaluate and test the efficiencies of 16 major Chinese commercial banks during the third round of the Chinese banking reform period (2003–2011) with the variable returns to scale setting and the consideration of undesirable/bad output. The main findings of this study are as follows: (i) the two-stage DEA model is more effective than the conventional black box DEA model in identifying the inefficiency of banking system, and the inefficiency of the Chinese banking system primarily results from the inefficiency of its deposit producing sub-process; (ii) the overall efficiency of the Chinese banking system improves over the study period because of the reform; (iii) the state-owned commercial banks (SOBs) appear to be more overall efficient than the joint-stock commercial banks (JSBs) only in the pre-reform period, and the efficiency difference between the SOBs and the JSBs is reduced over the post-reform period; (iv) the disposal of non-performing loans (NPLs) from the Chinese banking system in general explains its efficiency improvement, and the joint-equity reform of the SOBs specifically increases their efficiencies.     

Evaluation of bank branch growth potential using data envelopment analysis

Banks occasionally employ frontier efficiency analyses to objectively identify best practices within their organizations. Amongst such methods, Data Envelopment Analysis (DEA) was found to be one of the leading approaches. DEA has been successfully applied in many bank branch performance evaluations using traditional intermediation, profitability and production approaches. However, there has been little focus on assessing the growth potential of individual branches.This research presents five models that examine three perspectives of branch growth. Each model was applied to the branch network of one of Canada׳s top five banks to gauge the growth potential of individual branches and to provide tailored improvement recommendations. Using various analysis methodologies, the results of each model were examined and their functionality assessed.     

中国金融对经济增长贡献的效率研究——金融发展的阶段划分及评价

通过聚类分析将改革开放以来的中国金融-经济系统分为传统金融时期、金融转型初期、金融转型中期三个阶段,然后选取资本存量、劳动力、金融机构人民币资金来源、M2、贷款余额作为投入变量,GDP作为产出变量,采用非径向超效率分析模型测度中国金融效率。研究表明:中国的金融-经济系统存在结构变化,金融的绝对效率较低,但相对效率水平较高,中国已经顺利完成了金融体系的转型。     

房地产上市公司的绩效：基于DEA的2008年报数据分析

2008年对于中国的房地产上市公司而言是面临严峻考验的一年，房地产上市公司在面临来自国家宏观调控的同时又面临金融危机带来的楼市、股市的考验。如何面对上述考验，文章从定量研究的角度，利用数据包络分析（Data Envelopment Analysis简称DEA）对中国房地产测评中心于2009年5月19日在上海发布的《2009年中国房地产上市公司测评研究报告》中的《2009年沪深房地产上市公司综合实力榜》的前二十名企业的绩效进行评价，DEA的方法可以同时考虑多投入、多产出的情况，因而能够对房地产上市公司的绩效做出有效评价，并对效率相对低的公司做出建议。     

带权重约束的DEA超效率评价模型及其应用研究

针对产学研协同创新项目的相对效率比较问题,数据包络分析（DEA）法是比较有效的数量分析方法。但DEA法所构建的优化模型中,由于约束条件没有充分考虑评价指标体系内部指标之间客观存在的内在权重大小信息,从而导致模型过分强调了评价单元的投入产出特征差异,甚至出现了原本非常重要的指标其权重却为0的不合理现象。本文提出DEMATEL法和数据包络分析法有机结合的思想,将运用DEMATEL法获得的指标之间内在的权重大小信息,纳入DEA法非阿基米德无穷小量ε限制的CCR模型及其超效率模型的约束条件,创新模型的可行域,构建了权重约束型ε-CCR模型和超效率ε-CCR模型。并以美国2014-2015年度国家科学基金会产学研合作研究中心发布的年度报告数据,对美国71个以高校为主导的产学研协同创新中心的相对效率进行了实证分析。实证结果显示,权重约束型超效率ε-CCR模型比没有权重约束的超效率ε-CCR模型更有效,不仅能对多个评价单位进行全排序,而且能克服无权重约束的超效率ε-CCR模型存在的评价结果不合理现象。这些结果有助于进一步推动DEA法的理论创新和应用推广,并能为多投入多产出项目的决策提供新的方法,使实际问题决策更合理更科学。     

基于E-NSBM模型的中国省际资源环境效率动态评价

从资源环境系统的内部结构出发,将资源环境评价体系分为生产系统和环境治理系统,针对该结构建立了扩展强自由处置性下的网络DEA模型——E-NSBM模型,将其与窗口分析法相结合,对2005-2012年中国各省资源环境效率进行了评价。结果表明：我国经济发达地区的生产系统效率普遍较高而环境治理系统效率较低;中、西部地区环境治理系统效率较高,但东部地区环境治理系统效率改进幅度最大;综合效率表现差的省份普遍表现为高的能耗强度、碳排放强度和硫强度。因此,各省应根据各自发展需要来确定资源环境效率的改进途径。     

基于DEA的中国科普投入产出效率初步分析

文章介绍了中国科普投入产出的基本现状,结合中国科普工作的实际情况,构建了中国科普投入产出评价指标体系初步框架.应用数据包络分析法(DEA),选取了中国科普统计2006-2013年间的数据,分析评价了中国科普投入产出效率,得到效率有待提高的初步研究结论.     

京津冀三地创新绩效比较研究

京津冀地区是中国重要的经济区域,但近年来三地发展水平出现分化,京津两市与河北省的差距拉大,有出现“极化”效应的可能。通过构建创新效率的评价体系,对京津冀三地3年的技术创新投入和产出作DEA方法分析,结果表明河北省的创新效率明显低于京津两市。京津冀三地在综合发展水平、劳动力素质、R＆D活动经费投入和外商投资企业的技术溢出水平上的差异,是这种技术创新效率产生差异的原因。     

碳排放约束下能源经济效率的区域空间分布研究——基于能源消费结构门槛视角

利用2000—2012年样本数据对全要素省际能源经济效率进行了测算,研究发现样本期间绝大部分省市自治区能源经济效率整体都呈不断下降的趋势,并且生产过程中低效率地区大幅增加,分布状态由“点状”向“面状”扩散,尤其是中西部地区下降较为明显,究其原因主要为:二氧化碳排放和能源消费结构是能源经济效率变化和地区差异的重要影响因素;据此运用门槛分析方法对其进行区域空间分布研究,将全国30个省市自治区划分为低排放—结构较合理、低排放—结构欠合理、高排放—结构较合理类群、高排放—结构欠合理和高排放—结构不合理5大类群,最后结合各类型区域自身发展特点,提出因地制宜的低碳绿色发展建议,并从需求侧和供给侧两方面提出能源经济效率改善的政策建议。