Fellow Working Papers

Do women apply more for jobs when they know the hiring probability of female job seekers directly from employers? I implemented a randomized control trial and a double-incentivized resume rating to elicit the preferences of employers and job seekers for candidates and vacancies in Iraq. The treatment reveals the job offer rate for women, calculated using the employers’ selection of women divided by the total number of female candidates. After revealing the treatment, the women applied for jobs by three more percentage points than the men in the control group. This paper highlights the value of revealing employers’ preferences to improve the match between female candidates and employers when women underestimate the chances of finding a job. 

Alarming rates of forest loss in the Colombian Amazon have created a perceived trade-off that the only means of achieving economic prosperity is by sacrificing the forest. This study finds little evidence of this trade-off; rather, we find that economic development and forest protection are not an either-or choice. Forest clearing is driven by extensive cattle-ranching as a means to secure land titles. In essence, the loss of some of the world’s richest biodiversity is the result of some of the least economically complex activities that fail to achieve economic prosperity in the region. If anything, the acceleration in deforestation has accompanied a period of economic stagnation.

The existing economic model in the Amazon – centered on agrarian colonization and mineral extraction – has not generated prosperity for the people, all while failing the forest. The exceptional diversity of the Amazon’s biome is not reflected in the region’s economy. The Amazonian economy is best characterized by its low diversity and low complexity. A significant proportion of employment is linked to public administration – more than in other departments of the country. Very little of the production in the departments is destined to be consumed outside the departments ("exported").

This study seeks to define an alternative economic model for the Colombian Amazon from the perspective of economic complexity with environmental sustainability. Economic complexity research finds that the productive potential of places depends not only on the soil or natural resources, but on the productive capabilities—or knowhow—held by its people. This research finds that the Colombian Amazon will not become rich by adding value to its raw materials or by specializing in one economic activity. Rather, economic development is best described as a process of expanding the set of capabilities present to be able to produce a more diverse set of goods, of increasingly greater complexity. This model starts from the base of understanding the existing productive capabilities in Caquetá, Guaviare, and Putumayo, to identify high-potential economic sectors that build off those capabilities to achieve new, sustainable pathways to shared prosperity.

Achieving shared prosperity in the Amazon depends on the connectivity and opportunity in its urban areas. The primary drivers of greater economic complexity – and prosperity – are the cities in the Amazon. Even in the remote areas of the Amazon, the majority of people in Caquetá, Guaviare, and Putumayo live in urban areas. The low prosperity in the Colombian Amazon is driven by the lack of prosperous cities. The report finds that Amazonian cities are affected by the lack of connectivity to major Colombian cities that limit their ability to ‘export’ things outside the department to then expand the capacity to ‘import’ the things that are not produced locally as a means to improve well-being.

Did the COVID-19 pandemic create an opportunity to earn population control through illegal violence? We argue that criminal groups in Colombia portray as de facto police by using mass killings to reduce the COVID-19 outbreak. They used massacres as a threat to enforce social distance measures in places they considered worth decreasing mobility. Our results from an Augmented Synthetic Control Method model estimated that commuting to parks fell 20% more in areas with massacres than in places without mass killings. In addition, we do not find a decline in mobility to workplaces and COVID-19 deaths after the first mass killing. These findings are congruent with the hypothesis that illegal armed groups used fear to enforce mobility restrictions without hurting economic activities and their sources of revenue. However, violence slightly impacted the virus’ spread. Treated areas had a decline of 35 cases per 100,000 inhabitants four months after the first massacre.

This growth diagnostic report analyzes the economic constraints that explain the underperformance of the agriculture sector in Namibia. Section 1 starts by showing why Namibia’s agricultural challenge is unique when compared to the rest of the world. We then describe the sector’s key features, recent trajectory, and growth potential across different relevant dimensions in Section 2. In Section 3, we provide an adaptation of the growth diagnostic framework to the case of agriculture in Namibia and a detailed analysis of its economic constraints. Finally, Section 4 presents policy guidelines for addressing the challenges described in this report and prioritizing policy interventions accordingly.

We introduce incumbency effects into a tractable dynamic model of international trade. The framework nests the canonical Melitz (2003)-Chaney (2008) model as a special case. The key novelty is that fixed costs of market access decrease with tenure. As a consequence, there is less market exit and entry in response to a shock. We derive a gravity equation and show that, ceteris paribus, countries that liberalized their trade relationship earlier trade more today. We provide supporting evidence for the underlying mechanism and derive an augmented ACR formula (Arkolakis et al., 2012) for the gains from trade that accounts for incumbency effects. A quantitative analysis suggests that our mechanism can explain up to 25% of countries’ home shares and that the gains from trade are, on average, 10% larger when accounting for incumbency effects. The analysis further reveals novel distributional effects of trade that benefit real wages but reduce profits.

Tanzania’s manufacturing puzzles (and frustrations) seem to be a natural outcome of their policy choice. The Tanzanian economy experienced a significant acceleration over two decades, growing at a compounded annual growth rate of 6% between 1998 and 2018: Largest rates were recorded and sustained by the super commodity cycle (2005-2014). Within that growth trajectory, manufacturing’s share of gross domestic product (GDP) has lingered for 30 years below 10% – well below the 23% target established for 2025 in Tanzania’s Industrial Development Strategy (2011). As stressed by Diao et al (2021), the bulk of manufacturing value added is created by a few capital-intensive firms, whereas informal manufacturing has increased employment but without significant improvements in productivity/wages. Manufacturing exports surged in 2011 and remained steady since driven by subsector basic metals (gold & unrefined copper). If these are excluded, the curve mirrors the commodity price boom (likely a price boom rather than a volume boom). Looking only at exports conceals the fact that the bulk of the manufacturing output in Tanzania is sold in the domestic market rather than exported: exports are equivalent to less than 2% of GDP; domestic sales are seven times higher. While Food and Beverages make up for the largest share of manufacturing value employment and value-added, basic metals are the ones accounting for the vast majority of Tanzania’s exports.

The most binding constraint to investments in manufacturing in Tanzania is the availability and quality of electricity supply: Access to electricity is the lowest among peers, with large disparities between rural (22%) and urban (70%). Electrical outages are frequent and expensive for the manufacturing sector; firms even plan their production schedules and decide on plant locations based on power reliability. And yet, when we analyze the share of value-added against energy intensity at the sub-sector level, the negative relationship to be expected if electricity is indeed the constraint is there, but too fragile and noisy. Why? The strongest evidence points to the role of trade protection in compensating firms for other constraints, allowing existing manufacturers to capture large shares of domestic value-added while remaining uncompetitive in export markets. Large manufacturing subsectors of moderate to high energy intensity and more capital intensive enjoy higher tariff protection, creating a wedge that allows these industries to thrive in the domestic market. Despite joining numerous free trade agreements, Tanzania remains one of the most restrictive countries from a trade standpoint, eased by filing exceptions that shield individual products and entire domestic industries from competition. We have also found that effective taxation in Tanzania is relatively higher on labor (lower on capital, materialized through massive tax holidays granted within SEZ), skewing returns away from the country’s relative labor abundance. Failure to address the binding constraints creates a rationale for upholding protection, which reinforces biases towards capital and energy-intensive sectors. These policies go a long way in explaining the Tanzanian manufacturing puzzle.

We build a new longitudinal dataset of job tasks and technologies by transforming the U.S. Dictionary of Occupational Titles (DOT, 1939 -1991) and four books documenting occupational use of tools and technologies in the 1940s, into a database akin to, and comparable with its digital successor, the O*NET (1998 -today). After creating a single occupational classification stretching between 1939 and 2019, we connect all DOT waves and the decennial O*NET databases into a single dataset, and we connect these with the U.S. Decennial Census data at the level of 585 occupational groups. We use the new dataset to study how technology changed the gender pay gap in the United States since the 1940s. We find that computerization had two counteracting effects on the pay gap -it simultaneously reduced it by attracting more women into better-paying occupations, and increased it through higher returns to computer use among men. The first effect closed the pay gap by 3.3 pp, but the second increased it by 5.8 pp, leading to a net widening of the pay gap.

Can households' beliefs about future income shocks affect child labor? This paper examines whether the three-year gap between the announcement (in 2014) and the start (in 2017) of the Illicit Crop Substitution Program (ICSP) increased child labor in Colombia. The ICSP provides farmers with financial support for not planting and harvesting coca leaves – the key input of cocaine. My results from a difference-in-differences model using differences in historical coca production show that due to the ICSP announcement, children became four percentage points more likely to work in municipalities with historical coca production than in non–coca-growing areas. Although the likelihood of working increased in coca–growing areas, the hours worked per child declined modestly after the ICSP announcement. The expansion of the children working in coca fields but the decline in working hours per child produce null effects of the announcement on education outcomes. The rise in the expected income affects the time allocation decision within households in rural areas.

Combining satellite data with FAO potential yields we provide a new measure of South Africa's current and potential crop farming output. We find that field crop production is twice its census estimate, contributing 1.4% of GDP rather than 0.7%, and that achieving potential could increase its contribution a further 0.5% of GDP. Estimating horticulture potential is more difficult. We find that its 0.7% contribution to GDP is massively unreported, with actual production at 2.5%. Reaching potential could increase this number a further 0.5%. The distance from current to potential output represents over 100 billion 2017 rand of additional gross income and about 350.000 thousand jobs and is unevenly distributed across the country and concentrated in four provinces: Free State, Western Cape, Kwazulu-Natal and Eastern Cape. Our result suggests that there is room to expand agriculture, but because the potential gains are geographically concentrated, the solutions should have a strong location dimension.

Related project: Growth through Inclusion in South Africa

We discuss three cases of corporate-smallholder partnerships in South Africa’s former homelands, which have tried to bridge the problem of low productivity by supplying technology, technical assistance and financing along with established channels for sales and distribution. The cases are indicative of some key difficulties faced by such ventures: building trust, finding a suitable partner, successfully transferring technological to small farms, and reducing risk, particularly climate related. In order for these types of partnerships to help close the gap between South Africa’s two agricultures, solutions to these problems must be provided at greater scale. We explore mechanisms to achieve that scale, drawing lessons from South Africa’s successful franchising sector, as well as newly emerging business models and technologies from abroad.

Related project: Growth through Inclusion in South Africa

A menudo se piensa que alcanzar la prosperidad económica en la selva amazónica es incompatible con la protección del ambiente. Los investigadores ambientales suelen advertir, con razón, que la velocidad de la deforestación actual está llevando a la Amazonía a un potencial punto de quiebre a partir del cual la selva no podrá dejar de deteriorarse hasta convertirse en una sábana herbácea. Pero se habla menos de lo que hay que hacer para generar prosperidad compartida en las comunidades amazónicas. La deforestación suele tratarse como algo inevitable a la hora de atender las necesidades humanas, locales y globales. Este reporte sintetiza los hallazgos de dos proyectos del Laboratorio de Crecimiento de Harvard University, que estudian la naturaleza del crecimiento económico en dos contextos amazónicos: el departamento de Loreto, en Perú, y los departamentos de Caquetá, Guaviare y Putumayo, en Colombia. La meta de estas colaboraciones es valerse de la investigación de alcance global que ha hecho el Growth Lab sobre la naturaleza del crecimiento económico para aplicar esos métodos al reto único de desarrollar rutas hacia la prosperidad en la Amazonía, de manera que no se perjudique a la selva. Este reporte compara y contrasta los hallazgos en la Amazonía peruana y colombiana para evaluar hasta qué punto hay lecciones que se puedan generalizar sobre la relación entre crecimiento económico y protección del bosque en la Amazonía. 

The Colombian Amazon faces the dual challenge of low economic growth and high deforestation. High rates of deforestation in Colombia have led to a perceived trade-off between economic development and protecting the forest. However, we find little evidence of this trade-off: rising deforestation is not associated with higher economic growth. In fact, the forces of deforestation of some of the world’s most complex biodiversity are driven by some of the least complex economic activities, like cattle-ranching, whose subsistence-level incomes are unable to meet the economic ambitions for the region. All the while, the majority of the Amazonian departments’ population works in non-forested cities and towns, at a distance from the agriculture frontier that forms the “arc of deforestation.” The relative urbanization of the Amazonian departments, despite the vast land mass available, recognizes that prosperity is achieved through close social-economic interactions to expand the knowledge set available to be able to produce more, and more complex activities. Achieving economic goals therefore relies on creating new productive opportunities in non-forested, urban areas.

The risk of deforestation reduces incentives to improve the connectivity of Amazonian departments with major cities and export markets. The remoteness of these departments increases the cost of ‘exporting’ goods to markets outside the departments. Poor connectivity contributes to the low economic complexity of the departments. In turn, the low complexity reduces incentives to coordinate new investments that would generate returns to greater connectivity. Coordination failures, which occur when a group of economic actors (e.g., firms, workers) could achieve a better outcome but fail to do so because they do not coordinate their actions, are widespread in all three of the Amazonian departments studied. This limits the creation of new capabilities and productive diversification to generate new jobs and higher incomes.

We posit that economic growth in the Colombian Amazonian is limited by a “connectivity trap” whereby the lack of external market connectivity restricts economic complexity, and, in turn, the low complexity fosters the coordination failures that limit returns to new diversification. Ultimately, low returns to diversification further reduce incentives to improve connectivity. Underpinning the connectivity trap is the belief that limiting the connectivity of Amazonian departments with large Colombian cities and the broader global economy will limit incentives for deforestation. Yet, deforestation has accelerated in recent years, despite the continued poor connectivity. We argue that Colombia must create a new national law to curb deforestation by eliminating the financial incentives for land speculation. Reclassifying forested lands under the control of national protection systems with severe restrictions on economic activities and strengthened enforcement, as detailed in an accompanying report, provides the needed legal clarity regarding land formalization. Within the law to eliminate incentives for deforestation, the national government should create a new development approach for the Colombian Amazon. This approach must move beyond a natural resource-based approach to the region, to center on the productive potential of its urban areas, and the carbon markets and tourism potential of its forested areas. One pillar of this approach is to build new public sector capabilities to coordinate investments into new, targeted productive sectors to create new national-local mechanisms of investment promotion. A second pillar is to improve connectivity to external markets through road and air investments between Caquetá, Guaviare, and Putumayo and major cities and ports.

Las alarmantes tasas de pérdida de bosques en la Amazonia colombiana han creado la percepción de que el único medio para lograr la prosperidad económica es sacrificar el bosque. Este estudio encuentra poca evidencia de esta percepción; más bien, encontramos que el desarrollo económico y la protección de los bosques no son una opción entre uno u otro. La tala de bosques está impulsada por la ganadería extensiva como medio para asegurar títulos de propiedad de la tierra. En esencia, la pérdida de una de las biodiversidades más ricas del mundo es el resultado de algunas de las actividades económicamente menos complejas que no permiten lograr la prosperidad económica en la región. En todo caso, la aceleración de la deforestación ha ido acompañada de un período de estancamiento económico.

El modelo económico existente en la Amazonía—centrado en la colonización agraria y la extracción de minerales—no ha generado prosperidad para la gente y le ha fallado al bosque. La excepcional diversidad del bioma amazónico no se refleja en la economía de la región. La economía amazónica se caracteriza mejor por su baja diversidad y complejidad. Una proporción significativa del empleo está vinculada a la administración pública, más que en otros departamentos del país. Muy poca de la producción de los departamentos se destina a ser consumida fuera de los departamentos ("exportada").

Este estudio busca definir un modelo económico alternativo para la Amazonía colombiana desde la perspectiva de la complejidad económica con la sostenibilidad ambiental. La investigación sobre la complejidad económica encuentra que el potencial productivo de los lugares depende no sólo del suelo o los recursos naturales, sino también de las capacidades productivas (o conocimientos técnicos) de su gente. Esta investigación encuentra que la Amazonía colombiana no se enriquecerá agregando valor a sus materias primas o especializándose en una sola actividad económica. Más bien, el desarrollo económico se describe mejor como un proceso de expansión del conjunto de capacidades presentes para poder producir un conjunto cada vez más diverso y complejo. Este modelo parte de la base de comprender las capacidades productivas existentes en Caquetá, Guaviare y Putumayo, para identificar sectores económicos de alto potencial que aprovechen esas capacidades para lograr caminos nuevos y sostenibles hacia la prosperidad compartida.

Lograr una prosperidad compartida en la Amazonía depende de la conectividad y las oportunidades en sus áreas urbanas. Los principales impulsores de una mayor complejidad económica—y prosperidad— son las ciudades de la Amazonía. Incluso en las zonas remotas de la Amazonía, la mayoría de la población de Caquetá, Guaviare y Putumayo vive en zonas urbanas. La baja prosperidad en la Amazonía colombiana se debe a la falta de ciudades prósperas. El informe encuentra que las ciudades amazónicas se ven afectadas por la falta de conectividad con las principales ciudades colombianas que limitan su capacidad de 'exportar' cosas fuera del departamento para luego ampliar la capacidad de 'importar' las cosas que no se producen localmente como medio para mejorar el bienestar. 

La Amazonía colombiana enfrenta un desafío doble: bajo crecimiento económico y alta deforestación. Las altas tasas de deforestación en Colombia han llevado a que se crea que el desarrollo económico no puede tener lugar si se protege la selva. Nosotros no encontramos evidencia que sustente esa dicotomía: el aumento de la deforestación no está asociado a un mayor crecimiento económico. Las fuerzas detrás de la deforestación de una de las áreas con mayor biodiversidad en el planeta se sustentan en algunas de las actividades económicas menos complejas, como la ganadería extensiva, cuyos ingresos son incapaces de cumplir con las ambiciones económicas de la región. Al mismo tiempo, la mayoría de la población de los departamentos amazónicos trabaja en ciudades y pueblos desprovistos de selva, lejos de la frontera agropecuaria que forma el “arco de deforestación”. La relativa urbanización de los departamentos amazónicos, pese a la gran masa de tierra disponible, constituye un reconocimiento de que la prosperidad solo se logra mediante interacciones socioeconómicas que expanden el conjunto de conocimientos disponible para que se pueda producir más, y mediante actividades más complejas. Por lo tanto, para alcanzar las metas económicas hay que crear nuevas oportunidades productivas en las áreas urbanas sin selva.

El riesgo de deforestación reduce los incentivos para mejorar la conectividad de los departamentos amazónicos con las grandes ciudades y los mercados de exportación. El carácter remoto de estos departamentos aumenta el costo de “exportar” bienes a mercados que están fuera de estos departamentos. La conectividad precaria de la región contribuye a su baja complejidad económica, que a su vez reduce los incentivos para coordinar nuevas inversiones que podrían generar retornos a partir de una mayor conectividad. Las fallas de coordinación - que ocurren cuando un grupo de actores económicos (como empresas y trabajadores) podrían lograr un mejor resultado, pero no logran hacerlo pues no coordinan sus acciones respectivas - son extendidas en los tres departamentos amazónicos bajo estudio. Esto limita la creación de nuevas capacidades y la diversificación productiva que podrían generar nuevos empleos y mayores ingresos.

Planteamos que el crecimiento económico en la Amazonía colombiana está siendo limitado por una “trampa de conectividad” donde la falta de conectividad con los mercados externos restringe la complejidad económica, y a su vez la baja complejidad alienta las fallas de coordinación que limitan los retornos de una nueva diversificación. A fin de cuentas, los bajos retornos de la diversificación reducen aún más los incentivos para mejorar la conectividad. Como trasfondo de la trampa de conectividad está la creencia de que limitar la conectividad de los departamentos amazónicos con las grandes ciudades colombianas y el resto de la economía global limitará también los incentivos para la deforestación. Pero la deforestación se ha acelerado en los últimos años, mientras que la conectividad sigue siendo muy mala. Nosotros argumentamos que Colombia debe crear una nueva ley nacional para frenar la deforestación que elimine los incentivos financieros de la especulación con tierras, al reclasificar las áreas selváticas bajo control de los sistemas nacionales de protección para que tengan severas restricciones sobre las actividades que se puedan emprender en ellas, y se refuercen las labores de cumplimiento de la ley, como se comenta en detalle en el reporte siguiente. Con una ley que elimine los incentivos para la deforestación, el gobierno nacional debe crear un nuevo enfoque del desarrollo para la Amazonía colombiana. Este enfoque debe trascender el basado en los recursos naturales y centrarse en el potencial productivo de las áreas urbanas, así como en los mercados de carbono y el potencial turístico de las áreas selváticas. Un pilar de este enfoque es la construcción de nuevas capacidades en el sector público, que le permitan coordinar inversiones en nuevos sectores productivos específicos, para crear nuevos mecanismos locales y nacionales de promoción de inversiones. Un segundo pilar es la mejora de la conectividad con los mercados externos, mediante inversiones en carreteras y transporte aéreo entre Caquetá, Guaviare y Putumayo, y las grandes ciudades y los puertos.

Achieving economic prosperity in the Amazon rainforest is often seen as incompatible with protecting the forest. Environmental researchers rightly warn that rapid deforestation is pushing the Amazon close to a potential tipping point of forest dieback into grassy savanna. Less has been said about what is required to generate shared prosperity in Amazonian communities. Deforestation is often treated as inevitable to serve human needs, local and global. This report synthesizes the findings of two engagements by the Growth Lab at Harvard University that study the nature of economic growth in two Amazonian contexts: Loreto in Peru, and Caquetá, Guaviare, and Putumayo, in Colombia. The aim of these engagements is to leverage the Growth Lab's global research into the nature of economic growth to apply those methods to the unique challenge of developing paths to prosperity in the Amazon in ways that do not harm the forest. This report compares and contrasts the findings from the Peruvian and Colombian Amazon to assess the extent to which there are generalizable lessons on the relationship between economic growth and forest protection in the Amazon.

A growing body of research documents that the size and growth of an industry in a place depends on how much related activity is found there. This fact is commonly referred to as the "principle of relatedness." However, there is no consensus on why we observe the principle of relatedness, how best to determine which industries are related or how this empirical regularity can help inform local industrial policy. We perform a structured search over tens of thousands of specifications to identify robust – in terms of out-of-sample predictions – ways to determine how well industries fit the local economies of US cities. To do so, we use data that allow us to derive relatedness from observing which industries co-occur in the portfolios of establishments, firms, cities and countries. Different portfolios yield different relatedness matrices, each of which help predict the size and growth of local industries. However, our specification search not only identifes ways to improve the performance of such predictions, but also reveals new facts about the principle of relatedness and important trade-offs between predictive performance and interpretability of relatedness patterns. We use these insights to deepen our theoretical understanding of what underlies path-dependent development in cities and expand existing policy frameworks that rely on inter-industry relatedness analysis.

We develop novel measures of early-career skills that are more detailed, comprehensive, and labor-market-relevant than existing skill proxies. We exploit that skill requirements of apprenticeships in Germany are codified in state-approved, nationally standardized apprenticeship plans. These plans provide more than 13,000 different skills and the exact duration of learning each skill. Following workers over their careers in administrative data, we find that cognitive, social, and digital skills acquired during apprenticeship are highly – yet differently – rewarded. We also document rising returns to digital and social skills since the 1990s, with a more moderate increase in returns to cognitive skills.

South Africa’s labor market exhibits a unique equilibrium with one of the highest unemployment rates in the world and yet a low level of informal employment. The unemployment rate has remained high and persistent over recent decades, in spite of the formal demise of the apartheid regime and subsequent transition to democracy in 1994. This paper uses a matching model of the labor market to argue that spatial considerations combined with low productivity of informal work may be responsible for such an outcome. Spatial dispersion inherited from the apartheid regime thins the labor market, creating exclusion and perpetuating spatial segregation. In most developing countries, the result would be higher employment in informal or own account employment. However, with low productivity in the informal sector, the high rate of exclusion shows itself in higher unemployment rates instead. Transportation costs and housing deregulation may become key factors in improving the working of the labor market in South Africa especially if it is not possible to raise informal productivity.

Related project: Growth Through Inclusion in South Africa

Can South Africa’s segregation policies explain, at least partially, its current poor employment outcomes? To explore this question, we study the long-term impact of the forced resettlement of around 3.5 million black South Africans from their communities to the so-called “homelands” or “Bantustans”, between 1960 and 1991. Our empirical strategy exploits the variability in the magnitude of resettlements between communities. Two main findings. First, the magnitude of outgoing internal migrations was largest for districts close to former homelands. Second, districts close to former homelands have higher rates of non-employed population in 2011. Together the evidence suggests that districts that experienced racial segregation policies most intensely, as measured by outgoing forced resettlements, have worse current employment outcomes.

Related project: Growth Through Inclusion in South Africa

This report analyzes the economic legacy of spatial exclusion in South Africa, focusing on the long-term effects of the former Bantustan policy. Through quantitative analysis, the report explores the spatial dimension of economic activity in South Africa and specifically how this particular spatial institution has continued to shape current economic outcomes, despite past and present attempts to reverse the effect. The report also identifies areas for further research and potential intervention to enable more effective economic inclusion of the former homeland areas of the country.

Related project: Accelerating Growth Through Inclusion in South Africa