Productivity Improvements in LIC Manufacturing Sectors: Multidimensional Evidence from Bangladeshi Garment Factories

Manufacturing sectors in developing countries show a far greater dispersion in firm productivity and management practices than counterparts in high-income countries. Sustained increases in wages and employment creation depend on increases in productivity. This project aims to deepen our understanding of the process of productivity improvement in manufacturing firms in low-income countries (LICs) and to gain insights as to why productivity and innovation lag.

We focus on the Bangladeshi RMG sector, the largest source of urban employment in Bangladesh. We study the effects of a training program developed and offered by Solutions for Management International (S4Mi), a not-for-profit consulting firm established as a part of DFID's Responsible and Accountable Garment Sector (RAGS) program. The S4Mi program has elements of training and consulting, with both classroom and in-factory components. S4Mi and the participating factories select a single pilot production line on which to focus. During the first year of the program, there are seven one-week classroom training sessions, four focused on production / planning and three on HR. Each classroom session is followed by in-factory activations, which focus on providing tools and developing a culture suitable for problem solving. Following the initial activation phase, the program is rolled out to other lines in the factory. Further dimensions of particular interest, therefore, are the determinants of the diffusion of practices across the lines.

We incorporate a random selection of pilot lines to provide a valid comparison group prior to the factory-wide rollout. We combine extremely detailed productivity analysis with innovative survey and interview data. The productivity analysis provides an impact assessment of training provided as part of the project. Through the lens of the evaluation, the intense training offers an opportunity to observe a set of factories transitioning - or attempting to transition - internal processes and procedures. The survey and interview data will provide a unique window onto the internal processes of change that accompany the transition.

A sample of the key questions we address is:

a. Does the training program increase productivity, and, if so, do gains in productivity translate into better working conditions?
b. Which share of taught practices are managers aware of at baseline? Are managers capable of predicting which practices will lead to improvement and which won't?
c. Which factors affect diffusion of practices across lines? Do these factors differ depending on whether the suggested improvement comes from management or production floor workers?
d. Which production lines and manager characteristics correlate with resistance to change?

Mid-level management training is very commonly noted by key stakeholders as a critical bottleneck. Hence, there is demand for information about what works with regard to this training. We, the PIs, have been working on related projects in the sector in Bangladesh for more than two years. During this time, we have built important relationships with a number of foreign buyers (Tesco, Sainsbury's and H&M, for example), the industry associations (BGMEA and BKMEA), relevant government ministries (Commerce and Revenue), and foreign aid agencies (GIZ and DIFD). We will use these links both to inform the work during the project and to disseminate the results as we obtain them.

While the project fits squarely in the innovation theme, there is an important gender component as well. An estimated 80% of workers in the garment sector are female. Improving productivity in the sector, therefore, has the potential to have a tremendous impact on incomes of and job opportunities for women, as well as to reduce their working hours and to afford them more stable jobs.

Project beneficiaries include a range of stakeholders in garment sector in Bangladesh and other countries, including firms, workers, foreign buyers, foreign aid agencies, and policymakers. The unique data set and specific analysis will also benefit academics working on issues of productivity.

Impact will be maximized by involving key stakeholders in discussions from the beginning of the project and on a regular basis during the project. We, the PIs have developed relationships with numerous factories, several large foreign buyers, foreign aid agencies and high-level officials in relevant government ministries. We have regular discussions with these individuals related to work on other projects. The relationships and conversations will help ensure that the lessons learned are both relevant to the stakeholders and communicated to them. Foreign buyers and foreign aid agencies will also help to disseminate the lessons to stakeholders in the RMG sector in other low-income countries.

Stakeholders in the RMG sector in Bangladesh routinely point to the lack of skills in mid-level management as a key constraint in the sector. The efficiency level of the sector in Bangladesh - reported by various stakeholders and confirmed by detailed production data we have gathered in other projects - is typically between 35 and 40 percent. By contrast, efficiency is typically 70-80 percent in Sri Lanka. The industry thrives in Bangladesh despite this low efficiency because wage rates are much lower than competing countries. Increased wages cannot be sustained without increases in efficiency. This project examines one program for increasing productivity, but also tries to understand the constraints to increasing productivity.

The project will directly benefit about 300 mid-level managers in the RMG sector in Bangladesh, who will receive training. Learning whether, and how, training improves productivity will benefit not only the firms participating in the project, but other firms in the sector as well. We expect that at least part of any productivity gains arising from the training will be passed on to workers, 80 percent of whom are women, in the form of either higher wages or better working conditions. The RMG sector is the largest source of urban employment and the largest generator of foreign currency earning in Bangladesh.

The main academic beneficiaries of the project are researchers in the fields of development economics, organizational economics, industrial organization and international economics. Sociologists are another natural audience for the project, as are management scholars interested in the organization of international supply chains and issues related to social. Finally, the project is relevant for scholars working in areas of human capital formation, gender issues and female empowerment.

The project will produce and make available (in an anonymized form) an unprecedented data set. We will have detailed, line-level production from several hundred production lines in 30 factories covering a period of more than a year. Moreover, similar data gathered from related projects will make this the largest detailed data set for any collection of manufacturing firms in low-income countries. These data will be complemented with original data on trust and communication patterns within the factory. We will explore correlations between these variables and outcomes such as productivity, quality defects, lead time, retention rates, etc. Papers using these data will contribute to the recent revival in interest in the relationship between labour practices and productivity (see, e.g., Ichniowski and Shaw (2013)) by pioneering the study of the relationship between communication, trust and productivity. We expect these data will be useful for understanding many aspects of production which are unrelated to the training program we assess.