Nobody's business but my own: Self-employment and small enterprise in economic development☆
Introduction
Small businesses dominate the economic life of most developing countries. In Accra and Agra, Dhaka and Dakar, family firms and the self-employed account for the bulk of production and employment. This is true not only in agriculture and the service sector, but also in manufacturing. From cramped workshops and backyard foundries emerges an astonishing array of manufactured goods: clothing, footwear, pottery, metal products, processed foods, cement blocks, to name a few. In Ghana, as an illustration, more than 75 percent of the manufacturing workforce reports being self-employed, and fewer than 15 percent of manufacturing workers are employed in establishments with more than 10 workers (Republic of Ghana, 1987, Republic of Ghana, 1991).
In most rich countries, by contrast, small enterprises play a relatively minor role in economic activity—particularly in manufacturing. For example, in the United States, manufacturing establishments with fewer than five employees accounted for less than 1 percent of the value added in 1997, while firms with more than 500 employees accounted for almost half the value added (US Census Bureau, 2002).1 These data are consistent with a broad range of cross-section and time series evidence suggesting that as countries grow richer, small businesses and own-account work play a diminishing economic role.
What accounts for the differing importance of small firms and the self-employed in rich and poor countries? Can a neoclassical model adequately capture the relationship between economic development and the structure of production and employment? Is the small average firm size in poor countries necessarily the outcome of bad policy choices?
This paper attempts to shed light on such questions by analyzing a model that incorporates establishment size explicitly. The model, based on the Lucas (1978) span-of-control framework, is explored quantitatively, using parameters drawn from Japanese time-series data. The calibrated model suggests that the large differences observed across countries in establishment size and employment structure can be explained to a surprising extent by differences in productivity. Although distortionary policies—such as taxes that repress the growth of larger firms—undoubtedly play a role in exacerbating these effects, there would be substantial differences across countries even in the absence of distortions. Moreover, the model suggests that it is efficient in poor countries for many lower-skilled people to remain self-employed.
Section 2 briefly summarizes key facts concerning establishment size and economic growth, along with previous literature. Section 3 presents a dynamic general equilibrium model that is used to address the research questions. Section 4 describes the procedure by which the parameters of the model were chosen and the strategy for using the model to address research questions. Section 5 reports some results of interest, and Section 6 concludes.
Section snippets
Background and literature
As early as the classical economists, observers have noted that economic growth is accompanied by a concentration of production in ever-larger units and by a corresponding decline in self-employment and family enterprises. In more recent times, empirical work by Kuznets (1966), among others, documented this tendency in cross-country data. Kuznets suggested that one of the principal “characteristics of modern economic growth” was a series of shifts in the structure of production: from small to
A model of establishment size
To account for the observed abundance of small firms in poor countries, we need a model in which firm size is defined and in which productivity levels can vary. Standard models of neoclassical growth assume constant returns to scale at the level of the firm. As a result, these models do not address questions of establishment size.
Within the literature on industrial organization, there is a substantial body of theory on the nature of the firm and on firm size (see, for example, Coase, 1937;
Quantitative experiment
To compute solutions for the model, functional forms must be specified and parameter values assigned. For simplicity, this paper takes u(c)=log(c). For the production technology, it uses the standard CES form . Finally, the distribution for entrepreneurial ability, Δ(x), is taken to be a symmetric bell shape, which is modeled as a beta distribution, with parameters a=b=18.
The model parameters are chosen to match key features of Japanese time series data. During the 20th
Results
By construction, the calibrated model economy exactly replicates the entrepreneur–workforce ratios for the Japanese manufacturing sector in 1930 and 1992.6
Conclusions and implications
Previous theories of development have largely abstracted from questions of establishment size, despite substantial evidence that average establishment size—and particularly the level of self-employment—changes dramatically as economies grow. This paper suggests that a model with explicit treatment of establishment size and self-employment can reproduce a number of disparate features of the data. Not only can such a model mimic the data on entrepreneur–workforce ratios across a wide range of
References (38)
- et al.
Production, information costs, and economic organization
American Economic Review
(1972) - Biggs, Tyler, Pradeep Srivastava., 1996. Structural aspects of manufacturing in sub-Saharan Africa: findings from a...
Understanding Japan's savings rate: the reconstruction hypothesis
Federal Reserve Bank of Minneapolis Quarterly Review
(1989)The nature of the firm
Economica
(1937)The Other Path: The Invisible Revolution in the Third World
(1989)- et al.
An estimated model of entrepreneurial choice under liquidity constraints
Journal of Political Economy
(1989) Industrial structure and microenterprises in Africa
Journal of Developing Areas
(1994)- Gilchrist, Simon, John C. Williams., 2004. Transition dynamics in vintage capital models: explaining the postwar...
Getting income shares right
Journal of Political Economy
(2002)- Gollin, D., 2002b. Nobody's business but my own. Working Paper, Williams College Department of...
Property rights and the theory of the firm
Journal of Political Economy
Is Japan's savings rate high?
Federal Reserve Bank of Minneapolis Quarterly Review
The Strategy of Economic Development
The theory of the firm
Entry, exit, and firm dynamics in long-run equilibrium
Econometrica
Year-book of Labour Statistics 1940
Year Book of Labour Statistics Retrospective Edition on Population Censuses, 1945–89
Year Book of Labour Statistics
Cited by (125)
Why does the schooling gap close while the wage gap persists across country income comparisons?
2024, Journal of Economic Dynamics and ControlService on the rise, agriculture and manufacturing in decline: The labor market effects of high-speed rail services in Spain
2023, Transportation Research Part A: Policy and PracticeThe macroeconomic consequences of subsistence self-employment
2023, Journal of Monetary EconomicsSkill bias, financial frictions, and selection into entrepreneurship
2023, Journal of Development EconomicsCash transfers as a response to COVID-19: Experimental evidence from Kenya
2022, Journal of Development EconomicsCitation Excerpt :This population is one that is particularly affected by this crisis. In addition to making up the majority of employment in many developing countries (Gollin, 2008), these firms tend to operate in “non-essential” sectors and rely heavily on face-to-face interactions, leaving them vulnerable due to the particular features of the COVID-19 shock (Alfaro et al., 2020).1 In the first 4 months of 2020, average profit among our sample fell from 2 to 1 USD per day.
Structural change in labor supply and cross-country differences in hours worked
2022, Journal of Monetary Economics
- ☆
Early portions of this research were supported by National Science Foundation Grant NSF/SBR 9515256 and by the MacArthur Interdisciplinary Program for Peace and International Cooperation at the University of Minnesota. Subsequent research was conducted while I was a visiting fellow at the Economic Growth Center, Yale University; I greatly appreciate the recurring hospitality of the Growth Center faculty and staff. I am indebted to Tim Kehoe, Ed Prescott, Boyan Jovanovic, Jeffrey Campbell, Vincenzo Quadrini, Robert King, and Cheryl Doss for their comments and suggestions on early drafts of this manuscript. A working paper version of this paper, including details of the calibration procedure and sensitivity analysis, is available at: http://lanfiles.williams.edu/~dgollin/.