Purpose

Multi-directional efficiency analysis (MEA) is an alternative methodology to data envelopment analysis (DEA) that investigates the improvement potentials in each input and output dimension and identifies a benchmark proportional to these potential improvements. This results in a more nuanced picture of the sources of the inefficiency providing opportunities for additional conclusions about which variables the inefficiency is mainly located on. MEA provides insights into not only the level of the inefficiency but also the patterns within the inefficiency, i.e. its sources and location. This paper applies this methodology to Bangladeshi banks to understand the differences in the inefficiency patterns between different subgroups.

Design/methodology/approach

This paper analyses the difference in the pattern of inefficiency between the older family-dominated banks and the newer non-family-owned banks in Bangladesh using the recently developed MEAs technology, which enables analysis of patterns within inefficiencies rather than only levels of (in)efficiency. The empirical results show that whilst there are few significant differences in the levels of variable-specific efficiency scores between the two subgroups, there are clearer differences on the inefficiency contributions from particular outputs in most of the study period and also on most variables in the time window of 2007–2009. This finding provides clues to differences in business models and management practice between the two types of banks in Bangladesh.

Findings

The empirical results show that whilst there are few significant differences in the levels of variable-specific efficiency scores between the two subgroups (older family-dominated banks and the newer non-family-owned banks), there are clearer differences on the inefficiency contributions from particular outputs in most of the study period and also on most variables in the time window of 2007–2009, during the Global Financial Crisis (GFCs). This finding provides clues to differences in business models and management practice between the two types of banks in Bangladesh.

Practical implications

DEA is a conventional tool for benchmarking in management science. However, conventional benchmarking exercises based on DEA do not reveal significant differences in the sources of inefficiency that show differences in business models. While DEA remains the most utilized technique in the efficiency literature, we think that a more flexible and deeper analysis requires something like MEA.

Originality/value

The contribution is twofold. First, examination of performances of family-owned firms is a well-established but analysis of performances of family-dominated banks is relative scarce. Secondly, isolating the sources of inefficiency which differs between types of banks even if there is no difference in inefficiency levels is absolutely new for a complete data set of conventional banks in Bangladesh. It turns out that there are few (significant) differences between the groups in terms of the inefficiency levels, whereas clear patterns emerge in terms of differences in inefficiency contributions between family-dominated and non-family-owned banks, during the Global Financial Crisis

中文翻译：

---

孟加拉国家族银行的低效率模式

目的

多向效率分析 (MEA) 是数据包络分析 (DEA) 的替代方法，它调查每个输入和输出维度的改进潜力，并确定与这些潜在改进成比例的基准。这导致了对低效率来源的更细致的描述，提供了关于低效率主要位于哪些变量的额外结论的机会。MEA 不仅可以洞察低效率的程度，还可以洞察低效率的模式，即其来源和位置。本文将此方法应用于孟加拉国银行，以了解不同子集团之间低效率模式的差异。

设计/方法/方法

本文使用最近开发的 MEAs 技术分析了孟加拉国较老的家族主导银行与较新的非家族所有银行之间的低效率模式的差异，该技术能够分析低效率内的模式，而不仅仅是 (in)效率。实证结果表明，虽然两个亚组在特定变量的效率得分水平上几乎没有显着差异，但在大部分研究期间特定产出的低效率贡献以及该时间段的大多数变量上存在更明显的差异2007-2009 年的窗口期。这一发现为孟加拉国两类银行的商业模式和管理实践的差异提供了线索。

发现

实证结果表明，虽然两个子组（较老的家族主导银行和较新的非家族所有银行）在特定变量的效率得分水平上几乎没有显着差异，但从2007-2009 年全球金融危机 (GFC) 期间大部分研究期间的特定产出以及大多数变量。这一发现为孟加拉国两类银行的商业模式和管理实践的差异提供了线索。

实际影响

DEA 是一种传统的管理科学基准测试工具。然而，基于 DEA 的传统基准测试并没有显示出在显示商业模式差异的低效率来源方面的显着差异。虽然 DEA 仍然是效率文献中最常用的技术，但我们认为更灵活和更深入的分析需要像 MEA 这样的东西。

原创性/价值

贡献是双重的。首先，对家族企业业绩的考察是成熟的，但对家族主导银行业绩的分析相对较少。其次，即使在低效率水平没有差异的情况下，隔离不同类型银行的低效率来源对于孟加拉国传统银行的完整数据集来说绝对是新的。事实证明，在全球金融危机期间，各组之间在低效率水平方面几乎没有（显着）差异，而在家族主导和非家族所有银行之间的低效率贡献差异方面出现了明显的模式