From Capabilities to Opportunity: Driving SME Growth Christopher Woodruff, University of Warwick and PEDL Trade up and Compete Workshop Istanbul, Turkey May 11, 2015

Addressing constraints to SME growth Most research focuses on supply-side constraints – capabilities Entrepreneurship / management training: Do firm owners have the necessary skills? Finance: Do firms have access to the capital needed for expansion? Labor: Do firms have access to skilled labor?

Addressing constraints to SME growth Most research focuses on supply-side constraints – capabilities Entrepreneurship / management training: Do firm owners have the necessary skills? Finance: Do firms have access to the capital needed for expansion? Labor: Do firms have access to skilled labor?

Addressing constraints to SME growth Most research focuses on supply-side constraints – capabilities Entrepreneurship / management training: Do firm owners have the necessary skills? Finance: Do firms have access to the capital needed for expansion? Labor: Do firms have access to skilled labor?

Demand-side constraints These supply-side constraints may be very important. But is addressing the supply side sufficient? Is it always necessary? A smaller number of projects directly address the demand side. The results to date are promising.

Demand-side constraints Because I don’t want to overwhelm with information – and because there are not so many great examples yet – I’ll talk about three such projects. 1) An information intervention on product quality – private schools in Pakistan (Andrabi et al) 2) An export intervention among rug-makers in Egypt (Atkin et al / PEDL) 3) A formalization experiment in Sri Lanka (De Mel et al)

Changing market incentives There are several studies which aim to make markets more efficient, and hence change the incentives faced by firms. Look at a study of private schools in Pakistani villages. Basic idea: Provide information about product quality – test scores – and see how that affects behavior of schools. The intervention does not change the capabilities of either schools or teachers.

Pakistani school project design Sample of 112 villages; ~630 household per village average of 8 primary schools per village 166 children per village in school (grade 3) Some government schools, some private schools Timeline: Baseline survey 2004 – 1800 households testing early 2004, 12,000 children Report cards showing average results – English, math, Urdu – for each school in the village delivered to households in (56?) treatment villages in September 2004 retesting and resurveying early 2005 Question: Does the information on school quality induce underperforming schools to improve?

Pakistani school setup

How much did parents know? Even before the report cards, there is evidence parents knew something about relative school quality in their village A positive correlation between school-level perception of school quality (baseline survey) and average test score. A positive correlation between school fee level and: a) test scores; b) parent’s perceived quality But these correlations <<1. The test scores provide additional information

The effect of information about quality Average school fees fall in treatment villages by 17% Enrollment increases by 3.2 percentage points in treatment villages Average test scores increase by 0.11 standard deviations in treatment villages (relative to control) Weaker evidence for increased use of textbooks, less time spent on breaks.

Heterogeneity of learning effect

The effect of information about quality Average school fees fall in treatment villages by 17% Enrollment increases by 3.2 percentage points in treatment villages Average test scores increase by.11 standard deviations in treatment villages (relative to control) Weaker evidence for increased use of textbooks, less time spent on breaks. Further evidence of market effects: Scores improve most in the schools performing worst at baseline. Investment in Fees in the highest priced (and highest performing) schools fall.

Changing the “infrastructure” of markets The key point: The intervention was not designed to change capabilities directly. Yet product quality increased dramatically. Conclusion: Capabilities as “pushing the string.” The capabilities were there, but were dormant because the market did not provide the incentives to deploy them.

Changing the “infrastructure” of markets Casiburi, Glennerster and Suri (2014) examine the effects of rural road rehabilitation on market structure in Sierra Leon. (RD design rather than RCT.) Lowering transport costs has a significant effect on prices and entry by traders into rural markets Schoenholzer et al (in process) examine how the use of GSP / gyroscope / accelerometers on transport company minibuses affects contracts and quality of service (i.e., driver behavior) [PEDL-supported project.] Casiburi et al (2014) examine use of SMS messaging on productivity of contract farmers – changing incentives of the employees of the large contracting firm. [PEDL- supported project.] Several studies on the effect of credit bureaus on financial markets.

Providing a direct link to new markets: Exporting in Egypt Atkin, Khandelwal and Osman (2015) linked with an NGO Aid to establish links to export markets. Sample of 219 rug makers in Fowa, Egypt. Production for the domestic market

Rugs and technology

Providing a direct link to new markets: Exporting in Egypt With help of Aid to Artisans, identified a local intermediary and – eventually – buyers for rugs. Buyers in Germany, Israel, USA

Export markets

Providing a direct link to new markets: Exporting in Egypt With help of Aid to Artisans, identified a local intermediary and – eventually – buyers for rugs. Buyers in Germany, Israel, USA Randomly allocated the 219 producers to treatment and control. Gave production orders to the treatment weavers. A key point: The quality required for the export market is much higher than that required for the domestic market. The key question: Is there “learning by exporting”? Capabilities approach: Train the weavers to make higher quality products, then try to sell them in higher-income markets Demand-side approach: “If you build it, they will come”

Providing a direct link to new markets: Exporting in Egypt With help of Aid to Artisans, identified a local intermediary and – eventually – buyers for rugs. Buyers in Germany, Israel, USA Randomly allocated the 219 producers to treatment and control. Gave production orders to the treatment weavers. A key point: The quality required for the export market is much higher than that required for the domestic market. The key question: Is there “learning by exporting”? Capabilities approach: Train the weavers to make higher quality products, then try to sell them in higher-income markets Demand-side approach: “If you build demand, the capabilities will come”

Export orders Note: Slide from Atkin et al.

Did capabilities improve? Almost all firms (32 / 35) offered a “large” initial order accepted the export orders. Essentially all of them were able to deliver the product quality. Output – measured in square meters of carpets – decreased. And hours worked increased. But profits also increased.

Did capabilities improve? Almost all firms (32 / 35) offered a “large” initial order accepted the export orders. Essentially all of them were able to deliver the product quality. Output – measured in square meters of carpets – decreased. And hours worked increased. But profits also increased. Quality? A sharp increase in quality, judged in a blinded fashion by a carpet quality expert. For example, less waviness (lays flatter on the ground), squarer corners, closer to the design, etc.

Did capabilities improve? Almost all firms (32 / 35) offered a “large” initial order accepted the export orders. Essentially all of them were able to deliver the product quality. Output – measured in square meters of carpets – decreased. And hours worked increased. But profits also increased. Quality? A sharp increase in quality, judged in a blinded fashion by a carpet quality expert. For example, less waviness (lays flatter on the ground), squarer corners, closer to the design, etc. After adjusting for product quality, output increased among the exporters

Why does quality improve? First, it does not improve immediately, though very quickly. There is a learning curve. Note: Slide from Atkin et al.

Why does quality improve? First, it does not improve immediately, though very quickly. There is a learning curve. Further evidence of learning: The “quality lab” Atkin et al set up a shop and ask both treatment and control weavers to make a rug matching domestic standards. In spite of there being no reward to quality, the treatment weavers produce higher quality. How do they learn? Evidence from s suggests that the intermediary provides the training through feedback on the deliveries from the weavers. There is no evidence of actual investments, in equipment, consulting, etc.

Providing a direct link to new markets: Exporting in Egypt Almost all of a pool of existing small-scale manufacturers were able to transition to the higher quality standards required to produce products for export markets. Conclusion: Capabilities as “pushing the string.” The intermediaries increased the capabilities, but in a “learning by doing” fashion. The improvement in quality occurred quickly and without explicit investments.

Inducing entry of intermediaries: Other studies Burke and Falcao: Markets for maize in Kenya. An experiment to induce traders to enter randomly selected markets to understand effects of prices and incentives of farmers. [PEDL-supported project.]

Providing access to markets through formality Probably the most commonly examined issue on the opportunities side is formality. Does formalizing a firm open up opportunities to sell to large firms? To governments? An area in which PEDL has been quite active in supporting research

Incentivizing formality De Mel, McKenzie and Woodruff (2012): An experiment to encourage formality in Sri Lanka Take a sample of informal firms in Sri Lanka and randomize into 4 treatment groups and a control group: Treatment groups get information about registration, and then various levels of monetary incentives for registering. Follow-up survey months after formalization then used to measure whether and how formalizing benefits these firms.

Expected effect on informality Benefits of Formality Firms (CDF) Costs of Formality Formal Informal

What does ‘formalization’ mean? Dimensions of Formalization in Sri Lanka: Registration / compliance at several levels MC/UC/PS (local) License to operate Cost Rs annual license fee Main benefit is not being fined/harassed to register Division Secretariat Business Registration Certificate Cost if one-off fee of 1000 Rs + taxes if profits exceed 300,000 Rs per year Needed to sell to Govt, get Bank account in business name, serve as proof of legal existence of enterprise. District level VAT Employee benefits

Treatments Sorted into groups of 5 firms, matching on location, industry, # employees, willingness to register (61%), perceived benefits of being formal Treatment 1: Information on costs and benefits of being registered (prepared with DS, C of C), plus reimbursement of direct registration costs (modest) Treatment 2: Information plus payment of 10,000 Rs (~$85, or 1/3 rd of median monthly reported profit) Treatment 3: Information plus 20,000 Rs payment Treatment 4: Information plus 40,000 Rs payment Treatment 4a: 80,000 Rs.

Information Brochure

Open-ended direct questions to firms on what formalization has done for them:

Does formality matter? We find weakly significant increases in profits, and large but statistically insignificant effects on sales. But the effect is all in the upper tail of the distribution. A few firms grow a lot, most do not.

The upper tail In-depth interviews with firms suggest these increases real Automobile and autorickshaw firms who were able to become parts distributors after formalizing Saw mill able to put forest service stamp on receipt, allowing wood to get transported across municipal boundaries Grocery/tea shop – gone to get health inspection certificate to open bakery => May be quite a lot of heterogeneity in the channels through which impacts happen.

Formality as opportunity A very few firms respond to the opportunity which formalization creates by expanding. For the vast majority, business as usual. Conclusion: Capabilities as “preparation meets opportunity.” The increased opportunities from formality are enough only for a small percentage of the informal firms. For the majority, we conclude that either they lack something else or that formality is not an opportunity.

Inducing formality: Other studies Campos, Goldstein and McKenzie, encouraging formalization and providing information about banking services in Malawi. [PEDL-supported project.] Initial results indicate that combining information about formalization with information about financial services is much more effective. De Giorgi and Rahman, “carrot and stick” in Bangladesh. [PEDL-supported project.] Providing information alone was ineffective in increasing registration; “stick” treatment underway. Benhassine, McKenzie, Santini, and Pouliquen, encouraging registration for newly-created “entreprenant” status in Benin. [PEDL-supported project.]

Summing up For both the information on product quality (Pakistani school report cards) and the Egyptian rug exports, we see that producers demonstrate “hidden” capabilities when the incentives are changed to make it profitable for them to do so. For the informality project, the dormant capabilities appear to be present in only a small part of the sample. For others, perhaps standard capabilities are missing.

Concluding thoughts Understanding the constraints arising from a lack of capabilities is doubtless important. But that may not be the most important impediment in private sector development. Changing the structure of opportunities and incentives appears to be more important in driving growth in many contexts.

Thank You!