Medium-Run Impacts of Management Training in Garment Clusters

Abstract

This paper investigates the impact of management training programs on garment clusters in Vietnam and Tanzania. The study found that in the medium run firms showed improvement once they had identified useful practices and adapted them to their operations. Although it takes a few years to experience a significant impact on incomes, management training can increase not just management scores but also incomes or value added.

1. Introduction

The management score developed by Bloom and van Reenen (2007) and subsequent studies, along with several randomized controlled trials (RCTs) of management training, have confirmed a long-standing suspicion that management tends to be poor in the developing world. Hence, a question arises as to whether improvement in management increases incomes. In the survey of RCTs of management training, McKenzie and Woodruff (2014) point out that the evidence on this issue has so far been weak. Few studies find statistically significant impacts of experimental training programs on the value added of treated firms, and researchers have yet to arrive at a consensus on why the training impacts on incomes are limited. Are these poor outcomes the result of inadequately designed training programs, too early assessment of training impacts, or knowledge spillovers from training participants to nonparticipants?

Based on the RCTs of management training that were conducted in Vietnam and Tanzania (Higuchi, Nam, and Sonobe 2015; Higuchi, Mhede, and Sonobe 2019), this study attempts to provide a partial answer, particularly to the question of when the impact on value added materialized. By extending the follow-up observation period, this study found that although the treated firms adopted many management practices soon after the training, they stopped using some practices later. The initially insignificant training impacts on the value added, however, became significant in the medium run, suggesting that it takes some time for the treated firms to identify useful practices and modify them to fit their operations.

2. Study Sites

Since the ultimate goal of development economists is to prescribe an effective policy toward income generation in developing countries, it is important to evaluate training impacts in industrial clusters, which enjoy various benefits of agglomeration economies (Fujita, Krugman, and Venables 1999). Indeed, the vast majority of firms are located near other firms producing the same or related products (e.g., Sonobe and Otsuka 2011; Atkin et al. 2017). For the same reason, the study conducted impact evaluation in labor-intensive industries, in which developing countries are supposed to have comparative advantages. Since garment production is labor intensive and ubiquitous, garment clusters were chosen as study sites.

Conducting an RCT of management training in an industrial cluster has both advantages and disadvantages. A major advantage is that sample firms face the same prices of product, factors, and intermediate inputs, and have the same access to infrastructure because they produce the same products in the geographical proximity. This reduces heterogeneity among sample firms, thereby facilitating statistical inference.

A major disadvantage is that imitation is rampant in industrial clusters. Management practices and business performance might improve even for those firms that did not receive training, which would lead to an underestimation of training impacts unless a special method of impact evaluation, such as the one proposed by Baird et al. (2018), is applied. Although knowledge spillovers make impact evaluation difficult, the study notes that spillovers make the social benefit of training greater than the private benefit; this warrants further research on management training in industrial clusters. There is suggestive evidence for the existence of spillovers, however, the study did not apply any special method, and its results are likely to understate the impacts.

The garment cluster in Tanzania is located in Dar es Salaam, the country's largest city and has more than 700 garment producers, including the self-employed. The study randomly selected 113 firms out of the 250 members of three major associations of garment firms. They supply their products to the domestic market and occasionally export to neighboring countries of Eastern and Southern Africa region. In Vietnam, the study site is located on the outskirts of Hanoi, the capital city. It has numerous subcontractors, but they are not included in the sample. Instead, the sample covers all the 161 final-product producers in the cluster. They supply their products to the domestic market, and some of them export to Eastern Europe as well.

3. Experiments

Table 1 shows the sample size, the average number of employees, and other data on the sample firms. Typi-cal firms employ about 5 and 20 workers in the study sites in Tanzania and Vietnam, respectively. When a firm has no employees, what business owners must know about management includes self-management, financial management, and marketing. With many employees, owners need to know how to coordinate the division of labor as well. Thus, the experimental training programs covered not only basic accounting, marketing, and business strategy, as often adopted in the existing studies, but also elementary training in Kaizen management. Kaizen is an approach to production management and quality control, aimed at improving coordination among workers.

In both sites, the training programs consisted of two components: one offered classroom lectures for about 45 hours, and the other sent trainers to participants several times to provide coaching tailored to respective firms. In each site, the sample was randomly divided in half, and one-half was invited to participate in the classroom training component. Then, independently of this, the sample was randomly divided in half again, and one-half was invited to the on-site training component. Those firms that were invited to either one component or both were the treatment group, and those that were not invited to either component were the control group. The two groups in each study site differ in the baseline average firm size in terms of value added due to some outliers and the small sample sizes, but the difference is not statistically significant.

After the training programs were implemented in 2010, follow-up surveys were conducted twice in Vietnam and thrice in Tanzania from early 2011 through early 2014. As shown in table 1, the incidence of sample attrition was low probably because the sample firms were sufficiently large to survive and because even the control group continued to cooperate with the repeated surveys, expecting to participate in an advanced training program that is planned to take place in the future.

4. Major Results

An outcome variable of interest is the management score, or the number of good practices adopted by a firm. The score was constructed based on enumerators’ visual inspection and personal interviews with the owners of the sample firms.¹ In both sites, as shown in table 1, the treatment and control groups share about the same scores in the baseline, but their scores diverged from the first follow-up survey onwards (see the p-values for the difference in the changes from the baseline level). The control group's average score increased from the baseline through the second follow-up survey, suggesting that there were knowledge spillovers from the treatment group to the control group.² In Tanzania, the management scores at the third follow-up survey were lower than at the second, indicating that the firms stopped using some practices that they had adopted earlier. Still, the p-value shows that the difference in changes from the baseline is significant.

Another interesting variable is the annual value added, which is defined as sales revenue minus the costs of materials, electricity, and other intermediate inputs, and subcontracting costs. In both sites, the product markets were worsening throughout the post-training period, and the average value added declined substantially for the control group. In Tanzania, it continued to decline even for the treatment group although the magnitude of decline is smaller among the treated firms. Table 1 shows that the difference in changes from the baseline level was initially small and insignificant, but it became significant at the 5 percent level in 2012 in Vietnam and in 2013 in Tanzania. The results of more rigorous evaluation are reported in Higuchi, Nam, and Sonobe (2015) and Higuchi, Mhede, and Sonobe (2019).

The bottom two rows of table 1 show the fraction of the firms that were willing to pay the local currency equivalent of 150 USD for training participation. The fraction was very low in Vietnam, indicating that few business owners in this cluster knew the value of learning about management. Consistently, there were many firms that were invited to the training program free of charge but did not participate in it. In Tanzania, the fraction of willing firms was much higher and reached the upper bound after the training intervention. Indeed, the take-up rate of the training programs was nearly 100 percent. The firms in this cluster were more willing to learn probably because some successful entrepreneurs had started their businesses after participating in business training programs provided by international organizations and NGOs. In addition, the increased willingness to pay in the control group suggests that the training was favorably received by nonparticipants due to the knowledge spillovers.

5. Conclusion

This study presents new evidence that while management training can increase not just management scores but also incomes or value added, it takes a few years to experience a significant impact on incomes. The majority of firms are located in industrial clusters, where knowledge tends to spill over, and they may not know the value of learning about management. Such spillovers and ignorance which become a cause of market failure, together with the favorable training effects, suggest a need for policy intervention and hence warrant a considerable further compilation of research.

Footnotes

Notes

Yuki Higuchi (corresponding author) is an Associate Professor at Nagoya City University in Nagoya, Japan; his email address is [email protected]. Edwin Paul Mhede is the Commissioner General of Tanzania Revenue Authority (TRA) in Dar es Salaam, Tanzania; his email address is [email protected]. Vu Hoang Nam is an Associate Professor at the Foreign Trade University in Hanoi, Vietnam; his email address is [email protected]. Tetsushi Sonobe is Vice President of the National Graduate Institute for Policy Studies (GRIPS) in Tokyo, Japan; his email address is [email protected]. This work was supported by the World Bank Japan Professional Human Resource Development (PHRD) Trust Fund [grant number TF096317], the Japan Society for the Promotion of Science (JSPS) KAKENHI [grant numbers 25101002, 15K21728, 15H06540], and Grant-in-Aid for Research in Nagoya City University.

References