R4D Research Agenda
Round 1 Research Agenda (2014 - 2016)
The following are the detailed paper outlines for 2014-2016.
Paper 1: South African country study - a chronological overview of economic legislation/regulation and trends in performance of the South African economy and economic agents (firms and workers).
With the first democratic elections in South Africa, the newly elected ANC government inherited an economic system of mass unemployment, poverty and inequality with little worker protection. At the same time, South Africa was reintroduced into the global economy through the lifting of sanctions and deliberate trade liberalizations on the side of the South African government to welcome foreign investment. Against this background, the South African government formulated its economic policy with a particular focus on addressing the inequalities created by apartheid. However, it also created the environment in which South African firms had to develop strategies to balance increasing global competition and labour market regulations which were mainly directed at advancing worker rights rather than increasing employability.
This paper will follow a chronological approach and outline in the first part the development of economic policy frameworks and different policy tools implemented by the South African government from 1994 onwards. Starting briefly with the various macroeconomic frameworks over the time period from 1994 to 2014(RDP paper, GEAR, ASGISSA), we proceed to trade policy, industrial policy, fiscal and monetary policy followed by an account of the various labour market regulations and their policy instruments.
Against this backdrop, the second part of the paper provides a descriptive overview of economic trends, focusing particularly on the changing structure of production and employment in the South African economy. The paper starts by providing a picture of aggregate and industry/ sector specific economic growth trends and trade patterns linking these to sector characteristics such as productivity and employment intensity. It then draws upon a combination of econometric and decomposition techniques to identify the contribution of sectoral shifts in production on employment in South Africa, and the role played by international trade in driving these trends.
The analysis provides insight into the relevance of structural shifts away from labour intensive sectors as a source of the declining employment intensity of production in South Africa, as is argued by Rodrik (2008). Further, the paper builds on recent work by McMillan et al. (2013) on the role of structural shifts in employment in driving aggregate productivity growth in Africa. Finally, the paper extends and updates earlier research by Edwards (2001a, 2001b) and Jenkins (2008), whose Chenery-style decomposition techniques to separate out sources of change in production and employment into trade, demand and technology effects.
Most of the research done for this paper will be desk-top research. The first part will be based on a literature review tracking economic policies over the 1994-2014 period using government gazettes and official government information sources. The second part will make use of various data sets obtained from the South African Reserve Bank and the Statistical Agency of South Africa. Much of the empirical analysis will be conducted using industry level data at the 3-digit SIC level. To provide a more disaggregated account of changes in firms and the workforce, the paper will also draw upon various firm level data sets (including the Manufacturing Census 1996; Large Sample Surveys in 2001, 2005 and 2008; and the World Bank Investment Climate Assessment Surveys).
Paper 2: Continuity and Change: shifts and continuities in South African regulation of Labor Market since 1994 and the comparative analysis of the impact of selected labor market policies on employment. How has the Employment Equity Act of 1998 impacted on employment strategies of firms?
Since 1994, the legislative and regulatory environment of the South African labour market has been radically transformed, with a key focus of extending a large range of rights to all employees in order to address inequalities created under the apartheid regime (Benjamin, 2005). However, while the introduction of the new regulatory framework through the Labour Relations Act (LRA), the Basic Conditions of Employment Act (BCEA), and the Employment Equity (EE) Act has created a more secure work environment for some employees (especially in the formal sector), it has been argued that it has significantly increased the cost of employing labour. These cost, however, are presumably more burdensome for smaller firms and firms that are trying to compete in the global market (Rankin, 2006; CDE, 2013).
This paper explores the shifts in labour market regulation and legislation in South Africa from 1994 onwards, and investigates the impact of one particular labour market policy – the Employment Equity Act of 1996 - on employment and production strategies of South African firms.
The first part of the paper establishes a detailed account of the evolution of the various labour market regulations and policy instruments which have been implemented in South Africa since 1994. Each policy is contextualized in the political economy of South African interest groups and analyzed with regards to its intended goals particularly considering the trade-off between worker protection and labour market flexibility.
The second part of the paper provides an empirical investigation of firm responses to changes in labour market regulation. The Employment Equity Act (EE Act) of 1998 was designed to address unequal access to employment opportunities created by the racially segmented labour market under the apartheid regime. Thus, by legislating affirmative action, the Act requires firms that employ more than 50 employees to provide a detailed employment strategy over a five year period that outlines how the firm intends to restructure its workforce to reflect the demographic composition in the region in which it operates.
A number of recent international studies have investigated the impact of size dependent regulation on employment and production dynamics of firms around the policy assignment threshold (Garicano et al, 2013; Gourio & Roys, 2013; Ramasawy, 2013). A common finding is that the distribution of firms in countries with size dependent regulation exhibit a significant concentration of firms below the policy threshold which indicates that regulation influences firm decisions and firm growth.
This paper extends these analyses to the South African context. Using the fact that the Employment Equity Act is also a size dependent regulation which applies to firms of 50 employees and more, the paper proposes a Regression Discontinuity Design (RDD) approach to investigate the impact of the Employment Equity Act on firm dynamics for firms that fall just around the policy threshold. The research team is in the process of establishing a matched data set between firms that were interviewed in the 1996 Manufacturing Census and again in the 2001 Large Sample Survey (LSS) in order to track changes in employment intensity and production strategies of firms that operated in 1996 with a workforce of around 50 employees. This would allow us to observe firms before and after the implementation of the EE Act. If the standard RDD assumptions hold, then any differences in employment intensity or production strategies of policy treated firms that initially were above the threshold and therefore were exposed to the EE Act compared to firms that fell below the threshold can be attributed to the fact that the policy treated firms were required to comply with the Act.
The research team hypothesizes the following outcomes: between 1996 and 2001, firms that fell above the threshold had an incentive to reduce their employment numbers in order to fall below the threshold. Thus, we would expect firms above the threshold to have experienced a negative employment growth rate compared to firms that fell just below the threshold. Furthermore, while these firms might have reduced their employment numbers, they might have continued to produce the same amount of output by investing in more capital. This should also have been reflected in average wages as the more capital intensive production strategy would require a larger skilled/unskilled workforce ratio.
The similar analysis could be done on matched data of the 2005 and 2008 LSS. However, we should see almost the opposite trend for firms in the LSS data between 2005 and 2008. By 2005, firms had been aware of the policy and therefore had taken into account the additional cost and administrative burden of the Employment Equity Act. In such a case, we would hypothesize that firms that were above the 50 employee threshold in 2005were more likely to experience a larger employment growth rate compared to firms which were just below the policy threshold in 2005. Thus, firms just below the threshold had less of an incentive to grow their workforce which would have pushed them above the policy threshold compared to firms that already were operating above the policy threshold in 2005.
The above analysis will establish the average response of firms to the EE Act. It sets the groundwork for a more disaggregated study that explores how the EE Act mediated the employment response of firms in different sectors to different trade shocks. For example, Rodrik (1997) argues that trade liberalization raises the labour demand elasticity. We would, therefore, anticipate greater employment responses to the EE Act in firms exposed to international competition.
Paper-3: Anatomy of thriving and failing firms in terms of size, sector, employment strategy and investment behavior when facing changes in employment regulation
South Africa has a unique collective bargaining structure. Wages and working conditions, above a certain minimum set of conditions set out in the Basic Conditions of Employment Act, are mostly set at a sector level. Sectors where wages are set at a sector level are classified into two types: the first group where minimum wages are set by the Minister of Labour in consultation with a panel of experts; and the second where wages are set by Bargaining Councils generally comprised of the larger employers and employees in these businesses, and then extended automatically to other firms in the sector even if these firms were not party to the agreement. The first approach is used in sectors where worker organisation is difficult, such as agriculture, domestic workers and the retail sector. The second approach covers sectors like the manufacturing sector.
Bargaining Councils have been shown to cause lower levels of employment and fewer small firms in the areas and sectors where they exits (Magruder, 2012) and in sectors like clothing there is descriptive evidence of their deleterious effects (Nattrass and Seekings, 2012). However, these studies have only used cross-sections of data and broad definitions of coverage (Magruder, 2012) or focused on a specific industry (Nattrass and Seekings, 2012) and have not considered how they may constrain firms level adjustment, and consequently employment, to increasing import competition. The third paper will provide a detailed analysis of these Bargaining Councils and their relationship with wages and employment during a period of increasing trade in South Africa.
The third paper is the most data intensive paper in the three year period and requires that we build a database of Bargaining Councils and their decisions through extracting information from government gazettes. Wages and working conditions agreed upon by Bargaining Councils vary by coverage of the council, by region and by job description. This database will capture the various agreements of these councils and will link these to individual level outcomes in officially collected labour market data. The Post-Apartheid Labour Market Series (PALMS) combines 39 labour market surveys collected by Statistics South Africa since 1994, defines variables in a consistent way, weights observations consistently and calculates real earnings at an individual level during this period. Respondents in these surveys also report the size of the firms that they work for. This data will be linked to the Bargaining Council database through the sectors in which people work. We will thus be able to investigate sector level outcomes which are correlated to Bargaining Council coverage and agreements.
The paper will exploit three aspects of variation to identify the effect of Bargaining Councils on employment and wages: first, regional variation in coverage (Bargaining Councils set different wage levels between rural and urban areas, and some areas are covered by Bargaining Councils and some not); second, time variation (different Bargaining Councils have rounds of collective bargaining at different times); and third, variation in wage levels for different occupations within a Bargaining Council (different occupational levels have different set wages for different sectors and these do not correspond across sectors). This variation, and the changes which occur after different rounds of collective bargaining, can then be used to examine whether these agreements are associated with wage outcomes, employment levels and employment by different sizes of firms.
The results from this analysis can be used to discuss how this collective bargaining structure may constrain firm responses to increases in competition through imports. It seems likely that these structures will prevent firms from adjusting to increased competition, potentially from imports, by decreasing real wages. This in turn raises questions about how countries can create, or protect, better paying ‘decent’ jobs in a globalized environment.