Impacts of Kaizen Management Training: Experiments in Tanzania and Vietnam
Yuki Higuchi(Nagoya City University, Japan)
https://sites.google.com/site/yukihiguchipage/
Tetsushi Sonobe (National Graduate Institute for Policy Studies, Japan)
IEA Congress 2017 Mexico22 June, 2017
1
Background
• Entrepreneurs play a key role in industrial development and job creation
• Their managerial capacity is limited in developing countries
2
Management score developed by Bloom and van Reenen (2007 QJE)http://www.worldmanagementsurvey.com/
Background
• A number of RCT of management training
• Survey by McKenzie and Woodruff (2014 WBRO)
Significant training impact on managerial capacity
Impact on business performance or income is limited
-> Training content may be inappropriate
-> Impact is assessed too soon
3
What We Do
• RCT of management training featuring Kaizen
Kaizen: Japan-pioneered production management approach
Common-sense, low-cost, and human-friendly approach (not necessarily require capital investment)
-> It is likely to be accepted by SMEs in the developing world
• Targeting 100 and 300 SMEs in Tanzania and Vietnam
• Repeated follow-up surveys
4
• Kaizen is a basis of Toyota production system and origin of lean manufacturing 5
What We Find
In both countries, Kaizen management training
sustainably improved management capacity
improved business performance not in short-run but in medium-run
-> Changes in attitude: treated entrepreneurs started to learn management and/or became more proactive in business
-> Adaptive efforts to select useful management practices
6
Outline
Experimental design
Study site
Timeline
Intervention
Results
Next step
7
Study Site
• Industrial clusters: enterprises producing similar and related products locate near each other
Benefits of agglomeration economies (Fujita, Krugman, and Venables 1999)
Heterogeneity to be controlled is relatively small
Knowledge spillover is rampant (social benefit > private benefit)
-> We focus on three industrial clusters
8
Tanzania, garment
9
Vietnam, garment
10
Vietnam, steel
11
Basic statistics
12
Tanzania,garment
Vietnam,garment
Vietnam,steel
N 109 159 153
Years of education 10.7 8.1 6.8
Past training experience [=1 if yes]
0.64 0.13 0.03
Gender [=1 if female] 0.83 0.57 0.35
Baseline real sales revenue[1,000 USD]
22[13]
259[113]
1,767[1,197]
Baseline real value added[1,000 USD]
14[9]
75[29]
114[69]
Baseline number of employees5
[4]18[8]
20 [19]
Timeline
• Baseline survey (2010 Apr.)
• Classroom training (2010 Apr. - Sep.)
• Interim survey (2010)
• On-site training (2010 Nov. - 2011 Feb.)
• 1st follow-up survey (2011 Apr.)
• 2nd follow-up survey (2012 Sep. in TZ: 2013 Jan. in VN)
• 3rd follow-up survey (2014 Jan. in TZ: 2016 Jan. in VN)
13
TrainingJapanese Kaizen expert and local consultants with ILO qualification
Classroom training
• Lectures and workshop: 40 hours
• Production management as well as entrepreneurship, marketing and record keeping
• 45 / 50 in TZ and 93 / 197 in VN participated (ITT << TOT)
On-site training
• Instructors visited each enterprise: half day * several rounds
• Mostly production management
• 54 / 54 in TZ and 90 / 90 in VN received the consultation (ITT = TOT)
14
15
Vietnam, classroom training
16
Tanzania, on-site training
Sample Size
17
Group Classroom On-siteTanzania, garment
Vietnam,garment
Vietnam,steel
Class + Onsite
Invited Invited 26 32 32
Class-only
Invited Not 24 57 76
Onsite-only
Not Invited 28 16 10
Control Not Not 31 54 35
Total 109 159 153
Characteristics and baseline outcomes balanced
18
19
Results: management score (TZ)
20
Results: management score (VN)
21
Garment Steel
Management score
Kaizen score (Higuchi et al. 2015 JEBO)
• Information on adopted production management practices
• Based on 11 yes/no diagnostic criteria
• Enumerators’ visual inspection and/or entrepreneurs’ response
McKenzie and Woodruff (MW) score (2016 Management Science)
• Information on adopted marketing, procuring, record keeping, and financial planning practices
• Based on 26 yes/no diagnostic criteria
• Entrepreneurs’ response
22
Kaizen Score
23
Based on the enumerators' observationThe enterprise has a designated area for each production/activity within the workshop. The enterprise has a fixed place where major tools are stored. The storage of tools is put in order by kinds. The enterprise has a fixed place where raw materials are stored. The raw materials are stored separately from the scrap. The work flow line is determined. The defectives of raw materials and finished products are clearly segregated from good ones.
Based on the owners' responseThe scraps are removed and the floor is cleaned every day. The workers maintain machines every day. The enterprise holds meeting in which all workers participate. The proprietor knows how long each production process takes.
Results: management score
• Short-run impact on management score
• Medium-run impact on management score
• “Drop” suggests that entrepreneurs made adaptive efforts to select useful practices and to modify them to fit their business operation
24
Results: value added (TZ)
25
Results: unconditional value added (VN)
26
Results: survival (VN)
27
• 5 year survival rate: 68% in garment and 58% in steel cluster• Information from the exit enterprises (“missing” enterprises in the
3rd follow-up survey was only 5 in knitwear and 0 in steel cluster)
Results: business performance
• No short-run impact on business performance (value added, profit, sales revenue)
• Medium-run impact on business performance
• Particularly the combination of both components
• (In Vietnam) significant impact on five-year survival rate
28
Results: Willingness to pay (VN)
29
“Would you pay 150 USD to participate in the training program?”
-> “Are you definitely sure or probably sure?” (Blumenschein et al. 2008)
Vietnamese entrepreneurs, who has received parental support in business, did not know the importance of learning management
-> They became willing to learn after management
Garment SteelC+O C-only O-onlyControl C+O C-only O-onlyControl
Before training 28 16 38 11 13 8 20 3
After training 78 39 50 13 38 21 20 6
Real-world Behavior
Participation in NGO’s management training
• An international NGO provided management training in the knitwear cluster.
• According to our interview, they invited relatively large enterprises on mouth-to-mouth basis and did not systematically targeted (excluded) our training participants
• 17 (15 are treated) enterprises participated (significant at 10% in regression)
Product upgrading
• 17 (14 are treated) enterprises introduced new product by investing in machine and changing material supplier and sub-contractor (significant at 10% in regression)
30
Product Upgrading
31
Results are confirmed by regressions
• yit = outcome variable• Zi = 1 if invited to our training program (ITT)• yi0 = baseline value of outcome variable (if available) [McKenzie
(2012 JDE)]: ANCOVA specification• ηt = time dummies• εit = error term clustered at the enterprise-level
• We estimate LATE specification [Imbens and Angrist (1994 EMA)]: Replace Zi with Pi, which takes one 1 if participated in training program and use Zi as an instrument of Pi
• We check Lee bounds to take into account systematic attrition
32
Robustness check
• Trimming and winsorizing
• Controlling for record-keeping score
• Excluding distant recall data
33
Summary
• Kaizen management training sustainably improved management capacity of SMEs in Tanzania and Vietnam
• Training had impact on business performance, which took time to materialize
-> Longer-term evaluation is important
• The combination of classroom and on-site training worked
-> Kaizen component is most likely useful
• Demand-side constraint: lack of knowledge and spillover
• Supply-side constraint?: lack of reliable training provider
-> Governmental intervention warranted34
Next Step
Berge et al. (2012 J Af E)
• Management training for Tanzanian microentrepreneurs
• Training under close supervision of international researchers >> Training by local trainers
• Question about scalability
-> We are training Vietnamese youth as local consultants
• We document trainer training and their skill acquisition process
• The local consultants will provide training to local SMEs in future, which we evaluate with RCT design
35
Our team’s publication
• Mano, Iddrisu, Yoshino, and Sonobe (2012) World Development• Suzuki, Vu, and Sonobe (2014) Journal of Comparative Economics• Mano, Akoten, Yoshino, and Sonobe (2014) Journal of Japanese
and International Economies• Sonobe, Higuchi, and Otsuka (2014) Journal of International
Commerce, Economics and Policy• Sonobe and Otsuka (2014 Book) Cluster-Based Industrial
Development: KAIZEN Management for MSE Growth in Developing Countries
• Higuchi, Vu and Sonobe (2015) “Sustained impacts of Kaizentraining” Journal of Economic Behavior and Organization
• Higuchi, Mhede and Sonobe (2017) “Short- and Medium-Run Impacts of Management Training: An Experiment in Tanzania” under review
36