Diagnosing the barriers to retail sector productivity in Zambia

Economic policy, and the research that guides it, has often treated productivity as a single, amorphous feature that enables some firms to produce more output with fewer inputs. There are wide differences in so-called ‘total factor productivity’ (TFP) – that is, how efficiently outputs can be generated from a given amount of labour, capital, and other inputs – between firms even in a narrowly defined industry. But the underlying causes of those differences are unknown because the definition of TFP has no direct link to the firm's operations. 

Interventions targeted at low-productivity firms may not address the specific shortcomings of an individual firm, and may end up wasting resources or even causing harm. Much as doctors use diagnostic tests to help prescribe a drug, optimal interventions should be tailored to the limitations of each individual firm.

What drives productivity in Zambia's retail sector?

As part of an IGC-funded project in Zambia, we developed new statistical tools to understand productivity. These tools start from the assumption that a successful firm must excel in several distinct stages of production, and a manager who excels in one may struggle in another. The resulting statistics are each directly linked to a single stage – comparing them reveals the precise strengths and weaknesses of a firm’s management.

The project focuses on the retail sector, which employs millions of people across the developing countries – far more than manufacturing in nearly every country at every stage of development. We assume that retail has three stages of production, each with its own form of productivity:

1. Attracting customers: Holding fixed location, prices, and advertisement expenditure, a shop that attracts more customers has higher ‘marketing productivity’.
2. Sourcing and stocking inventory: Holding fixed labour and capital, a shop that transports and displays more inventory has higher ‘logistical productivity’.
3. Choosing the right mix of products: Holding total inventory and customer counts fixed, a shop that chooses high-demand products from low-cost suppliers has high ‘inventory choice productivity’.

Profits and management practices in the small-scale retail sector in Lusaka

We use these tools to understand the productivity of single-establishment retailers in Lusaka, Zambia. We surveyed firms in 25 retail markets using a detailed baseline questionnaire, followed by 8 months of high-frequency follow-up phone surveys. The sample includes shops at every level of sophistication, from stores with printed signs and paid employees to outdoor vegetable stands. There is also a wide variation in the gross and adjusted profits of these firms, as seen in Figure 1. Although the median firm earns roughly 500 Zambian Kwacha (ZMW) in adjusted profit, a significant share earns more than four times as much. 

Figure 1: Distribution of profits across firms in the sample

What makes some of these firms vastly more profitable than others? One explanation is the significant differences in how firms manage their marketing, logistics, and inventory. The most basic form of marketing, for example, is a sign displaying the name of the shop or advertising its products. Some shops have no sign or only a handwritten sign, while others post professionally printed signs. We find that firms with printed signs typically attract more customers than those without. 

Independent of marketing, firms that choose different (rarer) inventory also seem to earn higher profits. For instance, most shops selling onions will stock the white variety – but those that stock the red variety will, on average, sell more onions and earn greater profit on that variety.

Three factors affecting retail productivity

Our statistical tools capture productivity at each of these different stages of production, implicitly representing the skills necessary for distinct parts of the retail operation. We first explore whether managers who excel in one stage excel in the others. We find that the correlation between the different measures is low, suggesting that a univariate measure of TFP can obscure the relative strengths and weaknesses of firms if all three measures are independently important in determining overall profit. 

As a result, the new measures are jointly more informative about profit than existing measures of management or productivity. Comparing the new multidimensional tool to a set of standard yes/no measures of management practices (for example, “Does the shop keep written records of transactions?”) and a standard measure of total factor productivity, and found that the new multidimensional measures explain far more of the variation than existing methods.

How should we target interventions to support entrepreneurs?

Our results suggest that, when targeting and designing interventions to support the small-scale retail sector, governments and international development organisations should:

• Shift from general one-size-fits-all interventions to interventions that first identify a specific firm’s weaknesses (whether that’s marketing or logistics), and then provide support in that dimension.
• Assess a shop's flow of customers and skill at selecting inventory (say, by measuring the share of inventory that remains unsold) before giving inventory grants. The extra inventory may go unsold and be discarded, creating either direct costs (if the garbage must be collected) or public health risks (if garbage is left in the streets).
• Reevaluate programs that give physical capital to shops that are already skilled at logistics or unskilled at marketing or inventory choice, as they may respond by sourcing more inventory (again, leading to more waste). 

This blog is part of a series highlighting research supported by the Small and Growing Businesses (SGB) Evidence Fund, which generates evidence on the policies, programmes, and investments that can help firms grow, create jobs, and contribute to economic development.