Evaluating the impacts of WorkShop access for small-scale craftsmen in Kenya

Developing economies often have high numbers of low-capital enterprises that manufacture similar products and are located in close proximity to one another. While industrial clusters of very small firms provide some economic benefits, such as reducing input costs, firms operating in these clusters often operate inefficiently, use limited capital, do not consolidate or differentiate, and fail to grow significantly. Though various factors constrain the growth of small firms, several prominent impediments to growth are a lack of access to physical and human capital and access to markets.

A body of research indicates that these factors matter, and that alleviating constraints increases the productivity and profitability of small businesses. For example, in randomized controlled trials in Sri Lanka and Mexico, capital grants to small enterprises were very effective in boosting profits. The marginal return to capital in these studies ranged from 60% to 250%, with larger effects on more capital-constrained businesses. (McKenzie and Woodruff, 2008 ; de Mel et al., 2008). Even 5 years post-grant, treated Sri Lankan businesses had significantly higher profits relative to control groups and were more likely to have survived (de Mel et all., 2012).

Lack of necessary skills may also limit small business growth. Impact evaluations testing the effects of teaching business skills to entrepreneurs have generally shown moderate changes in adoption of business practices but uncertain effects on business outcomes (Karlan and Valdivia, 2010 ; Drexler et. al., 2011 ; McKenzie and Woodruff, 2012). The effects of non-business skills training have been more positive, with experiments showing large increases in participant wages after a variety of practical job skills trainings (Attanasio et al., 2011; Card et al., 2011)
Access to markets is another limiting factor for small businesses. Kenyan manufacturers we interviewed repeatedly referenced finding new customers as a main concern, expressed high levels of interest in marketing programs, and reported producing well below capacity. Evidence from other countries shows that enhanced market access drives improved quality and better performance for small businesses.

Though results from this research offer the promise of increased innovation, scale and success for small businesses, the translation of research into feasible and practical models for creating benefits has lagged behind. In particular, it is necessary to understand the appropriate combinations of support to small-scale entrepreneurs to maximize benefits and how to package and deliver that support in a financially sustainable manner that permits scale – this requires further experimentation.

To augment prior research we will evaluate the services offered by WorkShop – a social enterprise launching in Nairobi that is currently being incubated within the Busara Center for Behavioral Economics. WorkShop offers several services related to the constraints faced by small producers:

• Capital: WorkShop offers producers access to quality, industrial grade tools
• Human capital: WorkShop offers training (on design, production best practices and business management)
• Marketing: in addition to marketing training, WorkShop brands and markets products produced by independent producers

2017-06-05

2017-12-29

Approximately 3 months after craftsmen receive WorkShop services, Busara will conduct a follow up survey of all craftsmen. In addition to surveying craftsmen themselves, Busara will ask craftsmen to provide the contact information of recent customer and will survey those customers. Outcomes measured are described below. Outcomes with a * are primary outcomes, other outcomes are secondary outcomes.

• Production:
o Volumes by type of product (#)
o Productivity (index: relative productivity per type of furniture, defined as [pieces / hour] / [average sample pieces / hour], weighted by share of specific furniture type)*
o New designs created (#)*

• Enterprise outcomes:
o Enterprise revenues (KES)
o Enterprise marginal expenses (KES)
o Enterprise profits (KES)*
o Investment in capital goods for existing or new workshop (KES)*
o Employees (#)

• Labour:
o Hours worked (#)*
o Wages (if applicable) (KES / hour)*
o Job change in recent past (binary)

• Knowledge and practices:
o Adoption of practices included in training (sum of binary outcomes)*

• Customers
o Demographics: age, education, neighbourhood of residence, profession, approximate income
o Manner in which identified particular craftsmen (multiple choice)
o Other sources of furniture considered (multiple choice)
o Primary decision factors in purchase (1 – 5 scale)
o Overall satisfaction with piece purchased (1 – 5 scale)*

Initially, the Busara Center for Behavioral Economics will conduct a census of craftsmen in WorkShop’s target areas. This census will collect the contact information of craftsmen as well as basic demographic information and baseline data (e.g., baseline profits).

From the census data ~1,500 individuals will be identified as eligible to benefit from WorkShop’s services. As our goal is to improve the outcomes of small-scale craftsmen, we will only consider craftsmen associated with workshops having less than 5 individuals working there.

Due to capacity constraints, WorkShop is only able to provide services to a limited number of craftsmen. This creates the opportunity to randomize for the purposes of evaluation. Approximately 750 individuals will be randomly offered to participate in WorkShop’s core training program (design, production best practices and business skills). Randomization will be at the individual level and stratified by the size of the workshop the individual is associated with.

Within the treatment group of ~750 craftsmen, we will cross randomize access to physical capital – namely access to WorkShop’s tools and production facilities. ~300 randomly selected individuals will be invited to use WorkShop’s tools and production facilities. These craftsmen will be offered significant discounts on the cost of utilizing resources. Craftsmen in this group will access tools in waves of ~100 to accommodate space constraints.

WorkShop will also develop a small product line for which marketing, quality control and customer service support will be offered. The sample of craftsmen who can benefit from these services is dependent on customer demand. However, we will randomly allocate ~150 craftsmen to be “certified producers” who are eligible to benefit from these services. The actual numbers treated will depend on demand for products and we will use intent to treat (ITT) analysis to assess the impacts of this treatment arm.

Computer

Individual

No

1,500

1,500

750 - control
750 - training
300 - training and capital
150 - training, capital and marketing

Based on the sample sizes above, we are confident we can detect relatively modest changes in the variables above. We are able to detect a 0.13 standard deviation change in outcomes – generally considered a small change – for the overall test of the impact of access to WorkShop’s services. For the group receiving capital access or capital access and marketing, we can detect small to moderate changes in outcomes relative to the control group (0.17 and 0.22 standard deviations respectively). When comparing the treatment arms we can detect moderate effect sizes. Lack of uptake could impact the detectable effect sizes, but we expect high uptake of training and capital access given very high interest in receiving these services among our target population. Depending on market demand, we may not be able to create large opportunities for marketing services. There is a risk that we will fail to detect effects for this specific group but are a) confident in our ability to detect overall effects of WorkShop access and b) believe it is worth trying to measure the impact of marketing services.