The latest research on firms in Africa: A round-up of the 2018 Annual Bank Conference on Africa

This post is co-authored with Niklas Buehren.  Where possible, we indicate the method (e.g. #RCT).  A legend of the methods can be found at the end. 
 
Keynote
David McKenzie talked about the state of research on business training.   He took stock on six issues for this stream of research identified in McKenzie and Woodruff (2013). 

1. We need larger sample sizes:   We are seeing more studies with bigger samples but not for all because this won’t be possible everywhere
2. Improving measurement: triangulation is showing promise, RFID tags (don’t do this), photos of inventories (worked a bit better), verify business practices (working well).  
3. Does helping some firms hurt others?  Doesn’t look like it from McKenzie and Puerto.   Need to see more on this.
4. Do impacts of training differ over time?   Number of stories emerging here: a) catch up and failure (De Mel et. Al 2012, Brooks et. Al.), b) it takes time to see effects (McKenzie and Puerto 2017, Bruhn et al 2018) especially for employment, c) firms learn what works and knowledge spreads
5. What should be taught?   Promise from personal initiative, marketing and cost relative to standard curricula.   
6. Understanding market failures and solutions:  more work needed.  Ongoing work on understanding the price elasticity of demand for training, use the market for business service providers.

There are still knowledge gaps on these 6 areas plus:

• how do you scale up interventions?   (Getting the market to work better, use tech to reduce the cost and deliver at scale)
• how do we best target training programs?  
• other types of training for other types of firms (e.g. accelerators, export focused management improvements) 

And now, on to the sessions:
 
High Growth Firms

• Firm level characteristics are poor predictors of high-growth status. Removing distortions in product and factor markets would result in a two- to three-fold increase in the number of HGFs. Caution against targeting HGFs in a distorted environment since the level of distortion may be positively correlated with growth. #Cote d’Ivoire (Cirera et al.)
• Different measures/definitions identify different groups of firms as high growth firms (HGFs). Need for caution against the adoption of a single measure of HGFs, as each generates a sample of firms that are not necessarily appropriate for all forms of policy intervention. #RSA (Mamburu)
• Urbanization is linked to factor reallocation through firm turnover (entry and exit). Firm turnover contributes to the pace of factor reallocation and TFP growth. Transportation cost is a key predictor of this firm turnover. (Jones et al.)

 
Jobs, training and matching

• Business training increases mentors’ profits, but does nothing for their mentees profits. #Ethiopia #RCT (Bastian)
• Better connected people are more likely to get a job in a  #lab-in-the-field experiment in #Ethiopia.   Job referrals help overcome the exclusion of those who aren’t.   (Witte)   

Labor Markets

• Although supporting job applicants to land factory jobs provides substantial positive impacts on income, these jobs do not seem to live up to applicants’ expectations and may have adverse health impacts. #Ethiopia #RCT (Buehren)
• Paying for clicks gets news articles writers to produce fewer articles, spend more time per article, write about different things and reach a bigger audience. #Kenya #RCT (Tjernstroem)
• Reference letters can improve job search outcomes, in particular for those with weaker informal referral networks and improve firms’ ability to screen for higher ability candidates. #SA #RCT (Piraino)

Governance

• Distance to administrative HQs worsens access to public goods and economic development in remote areas which is partly driven by the quality of monitoring service provision. #India (Nagpal)
• Providing information and guidance on tax rules to firm owners, on average, reduces tax incidence but high-revenue firms are more likely to self-declare taxes. #Togo #RCT (Castaneda)
• Pre-Ebola investments into the health system led to increased health-seeking behavior measured by increased reported Ebola cases in targeted areas. #Sierra Leone #RCT (Christensen)

Financial Access, Credit, and Firm Performance

• Female entrepreneurs in Ethiopia show slightly more risk aversion than males.   Males seem to benefit from credit and training provided by the government, women don’t. #Matching (Yitbarek)
• Providing female entrepreneurs with a mobile savings/loan project results in increases savings and borrowing on the mobile platform.   The savings effects are higher when the mobile money is complemented by business training.   There is some evidence of crowding out of other sources of savings. Mobile money + training leads to some business diversification, but no increase in aggregate profit.   #RCT #Tanzania (Montalvao)  

Learning and Competition

• How to discipline network industries when there is no competition? In Rwanda, government intervention, by providing more licenses and greater competition in the mobile telecommunications industry leads to lower prices and increased aggregate welfare. #Rwanda (Björkegren)
• Are there spillovers from FDI to domestic firms? There is an 8% increase in the TFP of domestic firms in districts where FDIs operated. The knowledge transfers are mainly through labor flows, learning by observation/imitation and direct customer-supplier linkages. #Ethiopia (Abebe et al.)

 
Resource (Mis)allocation

• Removing distortions, such as corruption and tax rates, that drive misallocations could lead to TFP gains of 96% in #Ghana. (Ackah)
• Be careful what data you use but you’ll find that allocative distortions severely constrain TFP growth in Sub-Saharan Africa. (Jaef)
• Trade liberalization and tariff reduction improves resource allocation as well as it increases firm productivity especially for high-productivity and exporting firms. #Ethiopia (Zenebe)

Gender and Entrepreneurship

• In households of female entrepreneurs in #Ghana, income streams are managed independently and there is no expectation of transparency.  The different streams are earmarked for different types of expenses, although this doesn’t always happen in practice.  Women are trying to achieve three goals: reinforcing a partner’s responsibilities, b) fulfilling expectations, and c) long-term security.  This has implications for how they invest in their businesses.   #Qual (Pierotti)
• Are husbands of female entrepreneurs in #Ethiopia copreneurs, opponents, regulators or indifferent?   Mostly indifferent (especially when the husband is working long hours), followed by copreneurs (who experienced lower levels of household conflict) and regulators (who were more likely to hew to traditional gender roles).  #Latent Profile Analysis (Wolf)
• Do workers listen to female bosses?  Less than if she were a male.  But signals of her ability can overcome this.  Evidence from a signaling game.  #lab in the field #Ethiopia (Sheth)

Rural Enterprises and Livelihoods

• There are large yield gains to be had from changes in crop selection and composition of inputs. #Africa (Sinha and Xi)
• An asset transfer program that provides livestock and basic livelihoods training increases resilience among household participants. Beneficiaries are 44% less likely than control households to fall into poverty. #RCT #Zambia (Phadera et al.) 

Manufacturing and Growth

• Local market size is a major determinant of productivity in most industries in #Ethiopia.     #panel  data  #Ethiopia (Mengistae)
• A tour of growth in Africa (not just SSA!).   The last sustained period of economic expansion wasn’t due to commodities.  African industries that upgrade more disproportionately grow faster in countries that have better economic policies.  Of course, there is still room for improvement.  #diff-in-diff (Kiendrebeogo).    
• What determines productivity in #Ethiopia’s garment sector?  Labor and material inputs drive firm level outputs, with a weak effect from capital.  Human capital matters a lot. #Panel data  (Tekleselassie) 

 
Firm Productivity and Dynamics

• Learning-by-doing and self-selection leads exporting firms to higher productivity levels. #Ethiopia (Esmaile)
• Net entry rates of formal sector firms in #SA is declining in recent years. (Vukeya)
• Structural reforms lead to productivity improvements but the effect size depends on firms’ characteristics. (Kouame)

Energy and Climate

• Do higher temperatures reduce firm productivity? Yes, there are strong negative effects of temperature on firm revenues, TFP, profits and survival. (Traore and Foltz) 
• A study matching location specific barometer with enterprise surveys finds that living in a community with electricity outages leads to 35% lower chance of getting employment. The effect is significant for high skilled jobs. (Mensah)
• Electricity outages due to the power crisis of 2012-2015 in Ghana had large negative effects on productivity of manufacturing firms. There is a 10-12% reduction in productivity associated with electricity outages. #Ghana (Abeberese et al.) 

Labor markets and Training

• Expansion of vocational education reduces raises the unemployment of less educated workers. A 10% increase in graduates of vocational education reduces the probability of having a formal job by about 5% for less educated males. #Ethiopia #TVET (Fukunishi and Machikita)

 
Cities, Infrastructure and Firm Growth

• Are cities a source of productivity growth?    New agglomeration measures show that yes they do.   Moreover, increased connectivity is important for formal employment.   But, African cities aren’t generating the same benefits in terms of increased prod, wages, employment generation as cities in Latin America and Asia.   (Jones)
• How does infrastructure matter for trade reforms?   Road infrastructure magnifies the positive impact of input tariff reduction on productivity.  Evidence from #Ethiopia using an #IV (Sanfilippo)
• Efficient factor allocation increases firm survival, especially for marginal firms.  #Panel data (Nose)

 
 
Guide to method abbreviations
#RCT = randomized control trials
#diff-in-diff = difference in difference estimation
#Panel data = usually analysis of panel data without one of the methods above
#lab in the field = lab in the field experiments
#IV = instrumental variables
#matching = propensity score matching