Abstract
How come two developing countries with similar economic, institutional, and democratic attributes have developed vastly different welfare regimes? Drawing inspiration from the welfare regime literature, the author subjects Botswana and Mauritius to a comparative historical analysis that explores the economic and political trajectories of their welfare policy development. The findings offer both support and modifications to the established and mostly Western-oriented literature. First, politics affect welfare policy development. In developing countries, the rural population can promote welfare policy expansion, but it is middle-class interests that shape the direction of policy development. Second, given the importance of politics, conceptualizations of welfare policies need to include both the spending and the financing side. Finally, the causes of different welfare regime paths need to be examined with attention to the particular economic contexts of developing countries and how welfare and economic policies are interrelated and to some extent path dependent.
Keywords
Research on policy contexts and institutions of welfare has been remarkably ethnocentric. Mainstream welfare regime literature has almost solely focused on classifying and explaining welfare states and policies in OECD (Organisation for Economic Co-operation and Development) countries, with little regard to relevant policy experiences elsewhere (Walker & Wong, 2004). This is unfortunate as examining welfare policy experiences in countries of diverse political, economic, and cultural contexts may enhance our theoretical and empirical knowledge of welfare policy development.
There are some earlier examples of welfare policy research outside the OECD (Ahmed, Drèze, Hills, & Sen, 1991; MacPherson & Midgley, 1987). However, only recently has there been an upsurge in scholarly analyses on welfare regimes, welfare policies, and social security systems in the South. Much of this research focuses on middle-income countries (Bangura, 2007; Haggard & Kaufman, 2008; Sandbrook, Edelman, Heller, & Teichman, 2007), though there are also a few examples of studies including low-income countries (Adésínà, 2007; Gough & Wood, 2004; Townsend, 2009).
This article follows in the footsteps of these initial strides and thus contributes to the nascent literature on the causes of welfare regimes in developing countries. The aim is to explain the different historical trajectories of welfare policy expansion in Botswana and Mauritius, thereby strengthening our theoretical understanding of welfare policy development and expanding our empirical evidence. Specifically, by comparing two developing countries in a most-similar-system design, this article contributes to the welfare regime literature in three distinct ways: First, with few modifications, the power-resource approach is a valid explanation to welfare policy development outside the OECD. Peasants as much as urban labor are key advocates of welfare policy expansion, whereas the policy interests of the emerging middle classes are vital for the direction of welfare policy development. Second, it is critical to include in the analysis of welfare policy development the coinciding economic development and means of economic production. This is so not only because there tends to be a fit between welfare and economic regimes, but also because the distinct economic frameworks in developing countries affect the power resources of different classes and consequently the character of welfare policies. Third, this article suggests that the role of taxation should be included more firmly in welfare policy analyses. The tendency to exclude the financing side of welfare policies causes neglect of an important attribute of political interests and element in policy bargaining.
The article starts with a discussion of the key theoretical explanations of welfare policy development. This is followed by a methodological section that introduces the two cases, explains their suitability for fruitful comparison, and clarifies the approach used in the historical comparison. The main body of the article examines the causes of welfare policy development in Botswana and Mauritius in two steps: First, I focus on the formative years of policy development around and after independence, and, second, I concentrate on welfare policy expansion in the 1980s to 2000s and sum up by characterizing the countries’ distinct welfare regimes as they have developed. The conclusion reviews the causes of welfare policy development in Botswana and Mauritius and discusses the theoretical implications of these findings.
How Theories May Explain Welfare Policy Expansion in Developing Countries
In this section I present the main theoretical framework that steers the analysis. Given the wealth of welfare regime literature in the OECD, I draw inspiration from such scholarly contributions while I also benefit from the newer literature on developing countries. The theoretical argument has three interrelated elements: the role of class interests in policy bargaining (political alignments), the coinciding economic development (production regime), and the persistence of welfare policies once established (path dependence). Before elaborating, let me first clarify my understanding of welfare policies.
Welfare policies, as conceptualized here, consist of social benefits—social insurance and safety nets—and social services such as education and health (Haggard & Kaufman, 2008, pp. 3-4). Furthermore, though analytically playing a minor role, welfare policies also include the financing side, that is, those who are made to pay—mainly through taxation—for the social services and benefits.
Although some definitions of welfare policies include the state’s role in wage setting, employment, and general macro-economic steering (Esping-Andersen, 1990), such economic policies are excluded from the concept of welfare policies, but not from the general analysis. As I will elaborate on later, these economic policies, labeled by some as development strategies (Haggard & Kaufman, 2008, pp. 61-72) and by others as production regime (Huber & Stephens, 2001, p. 86), are underlying features that shape welfare policy development and that form part of the character of welfare regimes.
The term welfare regime indicates that welfare policies of different countries are organized around their own logic of stratification and social integration in such a manner that specific welfare regime types can be identified (Esping-Andersen, 1990). As we shall see, the welfare regimes of Botswana and Mauritius are distinctly different. Even so, it is not the purpose of this article to classify the welfare regimes of Botswana and Mauritius; rather, the aim is to explain how historical forces have caused such divergent welfare regime paths.
The Theoretical Model: Political Alignments, Production Regime, and Path Dependence
An important trait of the welfare policy concept defined above is the appreciation not only that policies tend to benefit certain social and economic groups—or classes—more than others but also that economic contributions to welfare policies may vary across classes. Understanding who will benefit and who is expected to pay the bill explains well who is likely to support welfare policy expansion and who will be more hesitant toward such developments. The power resource approach—a dominant explanatory concept in Western welfare regime theory—in large part follows such an understanding, even though taxation generally plays a minimal role in most welfare regime literature. From this perspective, welfare policy development is caused by the struggle between different classes over who stands to benefit from welfare policies (Esping-Andersen, 1990; Hicks & Swank, 1984; Huber & Stephens, 2001; Korpi, 1983).
The working class is seen as the main advocate for welfare policy expansion. Nevertheless, the course of expansion depends not only on the mobilization and organization of workers but also on the extent to which this class forges alliances with other classes. If the working class solely dominates policy making, it is likely that welfare policies will be more exclusively targeted toward this class. On the other hand, if the working class makes compromises with other groups, such as the middle class and/or high-income groups, welfare policies may have a broader reach. In fact, welfare policies that are broadly based and therefore widely supported will tend to lead to a more generous and encompassing welfare regime as everyone benefits and everyone pays (Esping-Andersen, 1990, pp. 16-33).
In the developing world, it is less straightforward to identify the traditional classes. However, the power resource approach is still expected to have merit. In their analysis of welfare state development in East Asia, Latin America and Eastern Europe, Haggard and Kaufman (2008) follow the power resource approach and analyze the role of political elites and how they incorporate or exclude the interests of urban labor and the rural poor. Like the Western-based power resource approach, Haggard and Kaufman bring to our attention the importance of interclass power struggles—critical realignments—for policy outcomes. Unlike the tendency to focus on the relationship between labor and capital (the economic elite), Haggard and Kaufman concentrate on the political elite. Moreover, they emphasize that peasants, not only urban laborers, are critical actors, and as such they are sensitive to the fact that in developing countries large parts of the populations still live in rural areas.
Undoubtedly, we must include the positions of political and economic elites as well as the organization of both urban and rural lower income groups in a power resource analysis of welfare policy development in the South. The analysis must also include the middle class—those between the elites and the lower income groups. Even though the group “in the middle” is heterogeneous, including small farm owners, white-collar employees, and small entrepreneurs in formal and informal sectors, these actors figure centrally in class coalitions (Sandbrook et al., 2007). In fact, the middle class’s commitment to welfare policies is crucial to a broad-based expansion of such policies (Esping-Andersen, 1990). Hence, although rural- and urban-based low-income groups are likely advocates of welfare policy expansion in developing countries, the trajectory of welfare policy development largely depends on the interests of the middle classes.
Based on the above discussion, I argue that to explain variations in the character of welfare policies in developing countries, it is critical to study political alignments. Political alignments refer to power resources and positions of different groups, that is, the organization and mobilization of classes as well as their scope of influence vis-à-vis other classes. The concept is similar to that developed by Haggard and Kaufman except that political alignments do not imply a discontinuity of power resources at a point of critical juncture such as a country’s independence. Rather, though the power resources of different groups may significantly shift because of specific events, the influence of organized interests may also be relatively steady, changing only slowly along with social and economic developments.
The actual interests of the main groups in welfare policy expansion are expected to be as follows: Economic elites (capitalists) promote policies that enhance their investment opportunities, such as low taxation and limited nonredistributive social transfers and services. Political elites are drawn by their need for political support; thus, their approach to welfare policies depends on who influences them, their need to make political compromises, and how potential policies affect their own livelihoods. Middle classes (urban and rural) are likely to support broad-based welfare policies if they stand to benefit; if the middle class as a whole gains little from policies, they are not likely to support expansion. Both urban workers and rural laborers support welfare policy expansion, particularly if the policies are redistributive; it is critical, however, that these groups are organized if they are to gain any political clout vis-à-vis the more resourceful elite groups.
It is then central to study political alignments; however, it is also important to include the economic context of developing countries in the analysis (Haggard & Kaufman, 2008). Economy plays a role in welfare policy development along two lines: First, the character of the underlying production regime—that is, economic policies and institutions (Huber & Stephens, 2001, p. 86) as well as the basic organization of the economy—determines the strength and behavior of various organized interests in society. For instance, high industrialization and progressive wage policies are likely to strengthen the position of urban labor. Second, the production regime not only shapes the organization of interests but also characterizes economic strategies. Just as specific production regimes tend to fit with certain types of welfare state in the OECD (Huber & Stephens, 2001, pp. 87-90), arguably so specific economic strategies coincide with welfare policy development in the South. In fact, welfare and economic policies—though theoretically separate—are often empirically interrelated and in combination define a specific welfare regime. As such, both welfare and economic policies may be part of the policy bargaining process. For instance, lower income groups may accept lower wages in return for some social security provisions, or the economic elite may agree to contribute to education and training against economic subsidies.
Both welfare and economic policies create feedback loops in the sense that the policies will tend to affect both the production regime and political alignment (Haggard & Kaufman, 2008, pp. 71-72). Policies on, for example, social security, social services, wage setting, and employment creation may affect both the way the economy is organized and the dominance of certain modes of production, just as such policies may benefit certain groups and arguably strengthen (or weaken) their position vis-à-vis other groups. In this sense, there is also a path dependence (Pierson, 2000) aspect to the theoretical argument proposed here. Once a certain type of welfare policies is in place, the cumulative commitments to such policies and the interests entailed therein will make it difficult to make dramatic changes to such policies. However, it is wrong to suggest that welfare policy expansion is locked in a self-reinforcing path. Rather, policy change is plausible, but it is a bounded change (Pierson, 2000, p. 265). Actors may introduce new policies or change old ones, but the decisions will be influenced by already existing policies as well as the broader political and economic context.
In short, and as illustrated in Figure 1, I regard the development of welfare policies as an outcome of a historical process where political alignments of actors are the immediate cause. The policy choices of organized interests are shaped by their scope of influence vis-à-vis other groups as well as by the underlying production regime characterizing the organization of the economy and economic strategies. Finally, as welfare policies develop through a process unfolding over time, initial policies may affect subsequent policy decisions, such that there is an element of path dependence in the character of the evolving welfare regime.
Model of theoretical argument
Comparing the Two “Miracles” of Africa
To test the theorized causes of welfare policy development two most-similar cases from the South have been selected. Botswana and Mauritius are appropriate for a historical comparison as they are reasonably similar on most competing explanations of welfare policy development while different on the main independent variable of interest: political alignments (Gerring, 2007; Lijphart, 1975).
The most common competing explanations of welfare policy development found in the literature refer to economic growth and trade openness; the rise of democracy, political parties, and governance institutions; demographics including population ageing and the extent of ethnic fractionalization; and external influences such as globalization and—for the developing world—the role of international agencies (Castles, 1998; Lindert, 2009; Prasad & Gerecke, 2010).
Botswana and Mauritius have separately been dubbed the “miracles” of Africa (Samatar, 1999; Subramanian & Roy, 2003). Certainly the two countries have fared well and generally outdo other African countries on many parameters; they are also remarkably alike on competing explanations, as Table 1 illustrates.
Comparison of Botswana and Mauritius
Human Development Report 2007/2008: Country Fact Sheets: Value is in US$2005, # measures rank among countries.
Economic Freedom of the World, 2006 data set (year 2004): Score from 10 (highest economic freedom) to 0 (lowest) # measures rank among countries.
The Economist Intelligence Unit’s index of democracy 2007: Score ranges between 10 (most democratic) and 0 (least democratic), # measures rank among countries.
Corruption Perception Index 2006: Score ranges between 10 (highly clean) and 0 (highly corrupt), # measures rank among countries.
Africa Development Indicators 2006 (World Bank): Measures on “voice and accountability,” “political stability,” “government effectiveness,” “regulatory quality,” “rule of law,” and “control of corruption.”
Botswana: Household Income and Expenditure Survey 2002/03; Mauritius: Household Budget Survey 2001/2002.
Development assistance as percentage of GDP, Human Development Report 2007/2008.
Botswana and Mauritius have capitalist, free, and open economies that have experienced long periods of sustained, high economic growth, and they are currently at a similar level of economic wealth. Indeed, both countries also score well, and quite similarly, on state capacity and good governance indicators, just as they are both rated as mature democracies with parliamentary systems and indirectly elected heads of states. Botswana and Mauritius have furthermore been subjected to external influences in similar ways; both have managed to negotiate beneficial access to the Western markets of their main agricultural products (sugar in Mauritius and beef in Botswana), and both have been reasonably free of internationally subscribed development strategies as their aid dependency has been limited. The two countries also have small populations and a similarly sized elderly population. Furthermore, even though Mauritius consists of more visibly diverse ethnic groups, the citizens of Botswana come from a variety of distinct tribal groups, where particularly the San people are economically and socially marginalized (Good, 2008). Even if Mauritius were considered to be ethnically more heterogeneous, the theory would then expect welfare policies to be more generous in Botswana than in Mauritius (Alesina, Glaeser, & Sacerdote, 2001), but as we will see in the next section, that is not the case.
Thus, Botswana and Mauritius have significantly similar economic, institutional, and democratic attributes. This allows for a gainful comparison as the distinct welfare regime paths of the two countries cannot therefore be ascribed to the alternative explanations of welfare policy development. At independence, the production regimes were not particularly different, as both countries were economically reliant on one main agricultural product; both were also relatively open economies that had small domestic markets, were remote from major export markets, and were vulnerable to exogenous shocks. For historical reasons, however, and critical for the analysis, the political alignments were (and are) significantly dissimilar: Mauritius, in contrast to Botswana, has long had strong and influential middle and lower-income classes. I illustrate how such differences have influenced the distinct economic and political trajectories with subsequent impacts on the nature of welfare policy development.
Comparative historical analyses are useful for testing theoretical arguments that suggest a complex interrelatedness of variables over a long-term period, as such analyses emphasize processes over time and make use of systematic and contextualized comparison (Mahoney & Rueschemeyer, 2003). The historical analysis is divided into two sections, and each examines the theoretical argument illustrated earlier: The first section, from prior to the countries’ independence to about 1980, describes the basic production regimes and political alignments and analyzes how welfare policies and coinciding economic strategies reflect the dominating political interests. The second section, from 1980 to 2008, equally starts with the production regimes and political alignments as they have developed and also considers how path dependence shapes a continuous welfare regime path. The analysis is based on careful examination of historical records, academic research, and policy documents and centers purely on the main critical economic, political, and policy developments in the two countries (see Figure 2 for a chronological overview of these).
Main developments in Botswana and Mauritius—chronological overview
The Early Years: Formation of Political Alignments and Policy Development
This first part of the historical analysis concentrates on the early years of policy development, starting before the countries’ independence to the early 1980s. We will see how the character of the underlying production regime shaped political alignments and how critical economic and welfare policies were in the interests of the influential political and economic classes.
Botswana: Modernization and the “Paternalistic” Elite
In the 1960s, as in previous decades, the Botswana economy was literally a cattle economy. Exports of cattle, beef, and livestock products accounted for more than 75% of total export earnings, and domestically cattle affected everybody’s lives. At least 90% of the entire population was rural, and the majority of Batswana owned some cattle—if not, they worked for and benefited from livestock owners. Even so, the national distribution of cattle was skewed, and at the same time, a relatively small number of farmers dominated the arable outputs (Colclough & McCarthy, 1980, pp. 23, 54-55, 110-127; Picard, 1987, pp. 114-115).
During the 1960s and 1970s, there were general increases in agricultural output and incomes because of good weather cycles, government inputs, and favorable beef prices. However, the economic turning point came in the early 1970s, when diamonds started to contribute to government revenues. Initially incomes from diamonds and other minerals were relatively small, but from the 1980s onward mining replaced agriculture as the most important economic sector. The ever-rising revenues gave the government opportunities to expand spending, and the period was characterized by high economic growth and rapid urbanization. The partnership between the government and the foreign-owned De Beers gave the diamond sector considerable influence (Colclough & McCarthy, 1980, pp. 57-99; Good, 2008).
It is well established that the political and economic elites were one and the same in Botswana. The politicians were well educated, and the majority of them were also wealthy cattle owners or businessmen. Of the members elected to parliament in 1969, all had 3 or more years of schooling, compared to the population, of which 90% had less, and 77% of the newly elected members owned more than 50 cattle (Hillbom, 2008, pp. 208-209; Parson, 1977, p. 642). Given their economic interests, politicians were committed to rapid economic growth within a system of free enterprise (Picard, 1987, p. 147).
As for opposition parties, these were generally very weak, and even if they did speak out against the government, they were also notably quiet on rural development issues, where they themselves had economic interests because of cattle ownership. Moreover, despite an increase in urban lower income groups during the 1970s and 1980s, the weak opposition parties failed to mobilize this class, and civil society organizations and trade unions were largely nonpolitical and controlled by government (Molutsi & Holm, 1990, pp. 329-330; Picard, 1987, p. 171). Contrary to opposition parties, the bureaucracy grew in influence as a political player as top management gradually became localized. The increasingly self-assured bureaucracy had self-interests related specifically to wages and benefits, whereas in other areas the preferences of civil servants were not unlike those of the political elite, as top-level bureaucrats were also keen to invest their salaries in cattle (Parsons, Henderson, & Tlou, 1995, p. 321).
The main political alignments were thus a political and economic elite united within the Botswana Democratic Party (BDP); but the bureaucracy also had increasingly strong influence over policy making, just as the diamond sector was another powerful, if less visible, partner. The rural poor—the majority of the population—were at the margin of the political decision-making process. In the elections in 1965, prior to independence, the BDP received overwhelming rural support, largely because of their ties to traditional leadership (Picard, 1987, p. 139). Subsequently, it was recognized that to maintain broad-based legitimacy, developments had to benefit all (Selolwane, 2007, p. 38), and the BDP projected an image of concern for rural development. On one hand then, there were still strong links between the elites and the rural areas, and rural support was essential to BDP (Parsons et al., 1995, p. 321). On the other hand, if preferences of the rural poor and cattle owners were contradictory, cattle owners would likely be given political priority (Colclough & McCarthy, 1980, p. 47).
This political framework influenced policy decisions. Though there was focus on nation building and broad-based development with infrastructural and social service expansion, social benefits and taxation received limited attention, and economic strategies included agricultural policies benefiting large cattle owners and wage negotiations, which, in the end, benefited upper-middle-class civil servants.
Initially, salary differentials in formal employment were narrowed considerably. However, from the late 1970s, as the influence of the powerful and articulate civil servants increased, previous efforts to reduce wage differentials were reversed and the main criterion in determining salary structures was no longer to reduce salary inequalities (Colclough & McCarthy, 1980, pp. 186-187). The main new agricultural policy was the Tribal Grazing Lands Policy (TGLP) of 1975, which gave exclusive access of commercial land to large cattle owners, enabling them to expand their beef production, whereas it became increasingly difficult for small livestock owners to keep cattle in overgrazed areas with only communal access to water. It is indicative of the situation that opposition parliamentarians were quiet on this issue and that top civil servants were also major beneficiaries of the policy (Holm, 1982; Picard, 1987).
Social service expansion was initially focused in urban areas, but eventually also expanded to rural areas as diamond revenue increased. Indicative of the interests of the political and economic elites, there was initially more focus on secondary and tertiary education as both the government and the mining industry were in dire need of skilled and qualified staff. Primary education received a boost in 1973, as did the health sector, when primary school fees were reduced (and abolished in 1980) and more funds were directed toward the service sectors (Colclough & McCarthy, 1980, pp. 210-214; Harvey & Lewis, 1990, p. 286).
Drought relief and feeding schemes were introduced around independence as early reactions to dire needs in this drought-prone and initially very poor country. Given BDP’s rural support base, it is no surprise that these schemes have persisted ever since—they have provided some (temporary and limited) safety net to virtually the entire rural population (Harvey & Lewis, 1990, pp. 59-60), and the drought relief program has been given as the single most important reason for BDP membership (Molutsi, 1989, p. 128).
Potentially redistributive policies such as taxation and transfers received limited attention. Symptomatically, for a government generally loath to use price subsidies, the cattle industry received substantial subsidies and tax exemptions; furthermore, as diamond revenues increased, the government continued to increase exemption limits and various tax brackets so that the nonmining income tax declined (Harvey & Lewis, 1990, pp. 263-265). The only social security policy introduced was the National Policy on Destitutes of 1980, which was initiated in a top-down manner as an administrative rather than legal document and merely provided a minimum of assistance to qualified destitutes (Bar-On, 1999).
Mauritius: Labor-Capital Compromises and the Political Middle Class
At independence Mauritius was, like Botswana, a mono-crop economy. Sugar production accounted for more than 90% of exports, almost half of all jobs, and more than a third of the gross national product. The little manufacturing in place was primitive and on a small scale, whereas the population and labor force grew rapidly, which caused increasing concerns about unemployment problems (Anker, Paratian, & Torres, 2001, p. 5; Meade, 1968, pp. 25-58). Despite such similarities, the political alignments, as they had formed decades before independence, were remarkably different from those in Botswana and influenced the direction of welfare policy making.
The sugar industry was characterized by great inequalities in ownership, where a small group of Franco-Mauritians owned the largest and most productive sugar estates. Many of these sugar barons had to sell off parts of their (less productive) land in the 1880s and 1890s because of economic difficulties, and although the Franco-Mauritians remained the economic elite, these developments gradually created a class of small landowners of mainly Indian descent. In some respects, the small landowners’ interests coincided with those of the larger sugar farmers, such as promotion of sugar exports and insurance of private property. On the other hand, there were strong ties to the farm workers as the laborers were also of Indian descent, and many small planters were, besides caring for their crop, similarly placed in wage-earning employment. The class of small planters furthermore prioritized education for their children, and over time this group became an educated and economically independent middle class with considerable political clout (Bowman, 1991, pp. 23-25; Bräutigam, 1999, p. 145; Houbert, 1981, pp. 79, 87-88; Simmons, 1982, pp. 6-10).
Even before independence, there was thus a strong, albeit small, capitalist economic elite of Franco-Mauritians. Indo-Mauritians constituted the largest group on the island (some 68%) and mainly worked as farm laborers, although there was an emerging Indian middle class. The Creole community, comprising people of mixed African, Indian, and European descent, was another relevant group. However, the Creoles were a heterogeneous group that could be found in nearly any economic positions—though the majority was urban laborers (Simmons, 1982, pp. 34, 44; Wellisz & Saw, 1993, pp. 222-223).
Two developments that shaped the political alignments in the 1930s and 1940s were critical to the early welfare policy initiatives. First, the Labour Party was founded in 1936 by the urban Creole middle class and Indian professionals and middle-class farmers who were able to mobilize lower-income groups (Bowman, 1991, pp. 30-31; Simmons, 1982, pp. 58-62; Wellisz & Saw, 1993, pp. 223-224). Second, in 1948 suffrage was extended such that it opened the door to more substantial Indian and Creole participation, whereas previously only wealthy Creoles and Franco-Mauritian had political influence. It is likely that the increasing voice of lower and middle income groups—particularly those in the sugar industry, which was aligned to the political Hindu class—pushed the Mauritian government to introduce a number of welfare policies such as a noncontributory means-tested pension for the old and the blind in 1950, from 1957 a universal noncontributory pension; cash benefits for needy persons; a Sugar Industry Pension Fund of 1956, which was a statutory, contributory earnings-related pension scheme for monthly paid workers in the sugar industry; and the Sugar Industry Labour Welfare Fund of 1948, which provided low-interest loans to sugar industry workers and their families. These programs were fairly modest (Bräutigam, 2008, p. 154; Mootoosamy, 1981, pp. 446-447; Simmons, 1982, p. 112). However, they set the path for future developments.
In the 1950s and 1960s, the political focus and negotiations concentrated on the status and constitution of a new independent Mauritius. However, it was also during this period that the Indo-Mauritian community consolidated its newly acquired political strength. The Labour Party shifted dramatically from being a class-based party working for the interests of lower income groups to becoming a Hindu-dominated party mobilizing agricultural laborers around Indian nationalism and Hindu culture (Simmons, 1982).
After the 1967 elections leading to independence, the Labour party created a coalition government “of national unity” with PMSD (Parti Mauricien Social-Démocrate), which was dominated by the Franco-Mauritian and Creole economic elites; this Labour-led coalition managed to stay in power until the early 1980s. It may be argued that this coalition was based on a concern for social harmony and national reconstruction (Kasenally, 2009, pp. 297-298). However, it is also worth remembering that Labour, as most parties in Mauritian political history, was led by middle-class people. This middle class of successful farmers, civil servants, and educated professionals, originating from the “landowning proletariat,” also had interests in a well-functioning sugar industry and adherence to private property (Houbert, 1981, pp. 87-88; Meisenhelder, 1997, pp. 279-280).
Hence, a strong Indo-Mauritian middle class with alliance to organized rural laborers dominated politically. And although segments of the Indo-Mauritians were increasingly consolidating themselves as the political elite, they were always distinct from the Franco-Mauritian economic elite, who were able to ensure some continued political influence, even though their interests were greatly compromised. Lower income groups were mobilized from early, on but they did not stay as one coherent voice for long. In fact, policy making was not dominated by an industrial urban class organized in strong trade unions—the group many welfare policy analysts otherwise often focus on. Instead, rural workers, organized and mobilized by middle-class sympathizers, composed the politicians’ electoral constituencies, and as such lower income groups held their influence in a labor-capital compromise controlled by an educated Hindu middle class.
Given the broad-based political alignments consisting of not only lower- and middle-income groups but also to some extent embracing the economic elite, it may not be surprising that the early welfare regime had an inclusive character. In general, welfare policies were not merely aimed toward providing relief for the poorest sections but deliberately aimed at achieving equitable distribution of wealth, whereas economic policies were directed toward creating productive jobs and establishing new investment opportunities for economic growth.
The range of welfare and economic policies—most of which still exist—were mainly introduced and/or extended during the period from around independence in 1968 through to the early 1980s. These policies include the comprehensive pension scheme, unemployment hardship relief, and free education. The expansion of social transfers and services in this period was financed, in the main, through taxation of the sugar industry in a steeply progressive fashion such that the large sugar estates paid most and the small farmers were exempt from taxation (Bräutigam, 2008, pp. 154-155; Wellisz & Saw, 1993, p. 239).
The welfare policies reached broadly, with some bias toward the laborers and small farmers in the sugar industry, whereas the economic elite generally provided the government with the necessary revenue. This deal was possible as economic policies provided the economic elite with incentives to compromise on their interests. The government followed a two-pronged industrial strategy: On one hand, it supported import-substituting industries, which catered to the small domestic market, and on the other, the government provided incentives to establish export-oriented industries for foreign markets. The export-processing zone (EPZ) promoted low-wage industries with few labor law requirements and provided preferential access to EU and U.S. markets. Combined with pervasive controls on exports of financial assets, it gave the economic elites plenty of incentives to invest their sugar profits in the EPZ industry (Bräutigam, 1999, pp. 147-152; Wellisz & Saw, 1993, pp. 232-234).
In the early 1970s, EPZ activities expanded, increasing employment opportunities dramatically; this coincided with the sugar boom caused by high sugar prices and record crops. Nonetheless, after 1976 fortunes changed: Sugar profits dried up, the world went into recession, growth in the EPZ industries declined, and the government’s budget deficits widened dramatically because of continued high expenditure and wage increases. By 1979, the Mauritian government was forced to turn to the IMF and the World Bank for economic assistance, initiating a period of structural adjustment programs (Lincoln, 2006; Simmons, 1982, pp. 193-194; Wellisz & Saw, 1993, p. 241).
Welfare Policy Expansion and Path Dependence
The previous analysis highlighted how the vastly different political alignments in Botswana and Mauritius affected economic strategies and welfare policies, but also how the underlying economic production regime—as it was initially and as it developed—shaped interests and policies. Hence, in Mauritius, the middle-class-initiated broad-based alliance called for policy compromises in regard to both financing and spending of welfare policy as well as economic strategies; these policy decisions in turn pushed the production regime toward a full-employment EPZ economy and, as we will see, strong support for broad-based welfare policies. In Botswana, to the contrary, the dominating political and economic elites focused on economic strategies, benefiting the cattle and mineral sectors, whereas the diamond-funded welfare policies focused on social service expansion. In the early years of Botswana’s development, the role of the middle class was limited. However, as evident in the following discussion, the growing middle class with few invested interests in redistributive welfare policies further determined the direction of Botswana’s welfare regime path.
Botswana: Reluctant Politicians and Indifferent Middle Class
Botswana followed a successful economic growth strategy relying on the creation of a reasonably open economy with state-led support for infrastructural and human capital development but also primarily intended to facilitate the mineral extraction industry and cattle sector. The economy was—and is—excessively dependent on the mining sector and foods and goods imports, with few structural changes in the patterns of production that would otherwise drive economic development and diversification (Hillbom, 2008).
In later years, as the elite grew in size and diversified their economic interests by investing in private enterprises and property development, their business connections and material dependence shifted away from the agricultural sector (Sebudubudu, 2009, p. 23). Still, the political and economic elite are in favor of policies promoting their economic interests and overall not in favor of redistribution and social security policies beneficial to lower income groups. Nevertheless, regardless of increasing urban attention, the rural poor remain loyal supporters of the political elite, whereas urban lower income groups are weakly represented in a split opposition and apolitical civil society organizations (Shale, 2009).
The emerging and predominantly urban middle class is likely to have few interests in redistributive policies. Given their position in relatively well-paid formal employment, they are insured against social and economic insecurity and may increasingly prefer private, employment-related solutions. Hence, this group may be reasonably content with the stability of the BDP government. Policy developments in the 1990s and 2000s is then continuously dominated by a strong, well-organized ruling party, whereas middle-income groups—who could potentially have interests in redistributive policies—have no natural alliances with lower income groups (Makgala, 2006, pp. 170-171; Taylor, 2005).
Many policies continued in the same form or with minor revisions during the 1980s to 2000s, with government thus sticking to the already-identified path. The few policy changes indicate an increasing shift by the upper and middle classes toward private solutions and a disregard for the need of any comprehensive social security legislation. The minimal social benefits were intended to reach the rural poor—still regarded as loyal supporters of BDP.
Although the opposition advocated for the introduction of comprehensive social security schemes from the 1980s (Hansard, 1988, p. 974), the BDP top leadership was in no rush to expand social transfers. In 1996, however, the BDP introduced a universal old-age pension—possibly because they were nervous about potential opposition inroads into the rural areas; certainly the pension scheme was a convenient instrument for the BDP to ensure continued rural support at a time of more even electoral competition (Good, 2008, pp. 53-54). In 2002, the destitute policy was revised in recognition of changed conditions since its introduction in 1980. In effect the revised policy changed little, but it is clear that it was politically important to secure minimal benefits for the destitute: “We are talking about our people and those are the most loyal people, destitute. Those are the people who wait from 5.30 am to 7 pm to cast their vote” (Mr. Butale, BDP, quoted in Hansard, 2002, p. 45). The trends in the taxation system—with a downscaling of income taxes and increases in indirect taxation—have continued. Minerals still account for, by far, the largest revenue source for the government, whereas nonmineral taxes, and hence also personal income taxes, are very minimal indeed (Good, 2008, p. 10; Mupimpila, 2005).
In the health and education sectors there are increasing demands for private facilities. Many higher- and middle-income groups make use of private schools, and private medical aid schemes are popular and prestigious and the number of beneficiaries is rising (Nthomang, 2007, pp. 198-199). Given these movements, it is not surprising that school fees for secondary schools were reintroduced in 2006. The middle and upper classes are likely to not have their children in public schools at the secondary level—and therefore would not be affected by the school fees—or, if they do, the amounts are bearable. For lower income groups, the cost sharing is more invasive and constitutes a move away from broad-based social services whereby all citizens benefit equally. Instead, the primarily diamond-funded welfare policies increasingly become irrelevant to upper- and middle-income groups, whereas the poor depend on the few meager benefits that are available—many of which have existed since independence.
Mauritius: Structural Adjustment Programs and Social Entitlements
In Mauritius, economic structural adjustment, economic growth, and employment creation characterized the 1980s. However, adjustments did not negatively affect the commitment to the established welfare provisions. The causes of such limited transformation can be found in the broad-based interests already invested in many welfare policies and the stability of political alignments; even though the Labour and PMSD coalition lost the 1982 elections to the rival MMM (Mouvement Militant Mauricien), Mauritian politics are now characterized by stability of policy orientation in successive governments (Kasenally, 2009; Srebrnik, 2002).
Trade liberalization was very gradual, and the state continued to provide subsidies and services to promote businesses as well as employment and training schemes for the unemployed. It is also worth adding that the already-established EPZ sector carried most of the boom from 1983 onward; the EPZ sector together with tourism grew rapidly, and employment increased dramatically to near exhaustion of labor reserves by the 1990s (Anker et al., 2001, p. 29; Wellisz & Saw, 1993, pp. 248-250). In the 2000s, because of pressure on the sugar and textile industries, Mauritius pursued economic diversification through ICT and offshore banking services to ensure continued economic growth and sufficient employment opportunities (Bundoo, 2006).
Mauritius has been able to retain its provisions of social transfers and to continue to deliver social services, despite IMF and World Bank recommendations to the contrary. The World Bank advocated that the Mauritian government change its role to a mere facilitator of economic opportunities and provider of safety nets to only the most vulnerable. Such recommendations attracted vociferous protests, and welfare policies continued to make up by far the highest portion of government expenditures in an economy that has overall been growing substantially over the past decades. When the government in 2004 introduced means testing on the universal noncontributory pension, it lost the elections the following year, and the new government reversed the means testing (Bunwaree, 2007, pp. 229-230; Meisenhelder, 1997, pp. 293-294; Sandbrook et al., 2007, pp. 127-129).
Such commitment to welfare policy benefits clearly reflects the legitimacy generated by the welfare regime. In fact, it has been argued that the right to social welfare is embedded in the psyche of the average Mauritian, including the politicians, and that it would be political suicide to suggest major welfare policy retrenchment (Bunwaree, 2007, p. 219).
It is interesting that the tax regime is the policy area most substantially changed since the early 1980s. With the help of IMF experts, Mauritius overhauled the entire tax system from the mid-1980s. Maximum income tax rates were reduced from 70% to a maximum of 30% in 1991, and corporate tax rates were equally reduced to 25% and 15% for manufacturing companies. In 2007, the government introduced a flat-rate personal and corporate income tax of 15%. Even though tax rates were reduced, revenues were maintained because of improvements in tax collections as well as the introduction of indirect taxes (African Development Bank Group, 2008; Anker et al., 2001, pp. 14, 77-79; Wellisz & Saw, 1993, p. 247).
Hence, the Mauritian welfare regime, as it has developed, is quite inclusive in that all have benefited from extensive social services, from social transfers, and through the creation of productive jobs and attractive investment opportunities. Historically, heavy taxation was laid primarily on the better-off, and even though the taxation system has changed, there appears to be a sense of contribution across the board. The case of Mauritius also suggests an interesting distinction in the inertial “stickiness” of welfare policies: Welfare policies on the spending side—where the population benefit—are hard to revert, whereas welfare policies on the revenue side—where the population contribute—are easier to retrench (and, in reverse, harder again to expand).
Welfare Regimes in Botswana and Mauritius anno 2008
The distinct welfare regimes in Botswana and Mauritius can largely be ascribed to the divergent historical experiences described above. The conclusion reviews the explanations for such divergent welfare policy paths, whereas this section briefly summarizes the character of the welfare regimes in the two countries anno 2008 before the economic recession set in.
As evident in Table 2, both countries spend a substantial part of the government budget on welfare policies in general, and both also prioritize education, health, and housing as a means to expand human capital. In Western welfare theory, the distinct typologies of welfare states are more often attributed to social security provisions rather than social services (Jensen, 2008), and it is also with respect to expenditures on social welfare that the two countries stand apart. Mauritius spends about one fifth of government expenditures on pensions, social security, and other social provisions, whereas such policies amounts to a meager 2% to 3% in Botswana.
Welfare Policy Expenditures in Botswana and Mauritius, Percentage of Total Government Expenditures
Source: Bank of Botswana, 2007; Ministry of Finance and Economic Empowerment, Mauritius.
Botswana: food and social welfare; includes the destitute persons program, vulnerable groups feeding program, orphan care, community home based care, primary school feeding program, old age pension, and World War II veterans grant. Mauritius: social security and welfare; includes various noncontributory and contributory pensions, child allowances, social aid, food aid, indoor relief, unemployment hardship’s relief, funeral grant.
Botswana: other community and social services; Mauritius: recreational, cultural, and religious services.
Within the social welfare category, both countries have a number of smaller, often means-tested programs, which are aimed at reaching vulnerable groups—that is, the destitute policy in Botswana and the social aid program in Mauritius. Mauritius, unlike Botswana, has an unemployment scheme, and especially its pension program is more extensive. The basic pension in Mauritius is larger (about double that in Botswana), and it is also paid out beginning at the age of 60, compared to 65 in Botswana. Furthermore, in Mauritius a pension is also available to the disabled, widows, and orphans, whereas in Botswana a pension refers purely to old age. Finally, the pension scheme in Mauritius is complemented by a well-developed contributory scheme, which tops off pension funds for many retired people (Seleka et al., 2007; Willmore, 2003).
As we saw, different economic strategies have accompanied the development of welfare policies, particularly with respect to employment creation and wage setting. Despite recent increases in unemployment rates in Mauritius to just less than 10%, Mauritians have better employment opportunities than do the citizens of Botswana, where official unemployment figures are above 20% (United Nations Development Programme, 2007, Table 21). Contrary to Botswana in later years, wage setting in the Mauritian public sector has explicitly promoted the idea of equity, which has ensured a reasonably equal wage development across the board (Bräutigam, 1997, p. 56).
Overall, therefore, both countries have emphasized human capital development as a part of their economic growth adventure, as evident in the focus on social services. Here, however, the similarities end. In Mauritius, policies have been more redistributive, highlighting income security and welfare through jobs and social security measures, and have been financed heavily through taxation (Sandbrook et al., 2007), whereas welfare policies in Botswana, financed mainly by diamond revenues, have played a minimal role outside social service provision (Nthomang, 2007).
Conclusion: Explaining Divergent Welfare Regime Paths
The comparison of Botswana and Mauritius highlights the role of political alignments in the formation and expansion of welfare policies. As such, this article offers theoretical and empirical insights that further enrich welfare regime research. Like Haggard and Kaufman (2008), the analysis discloses that the rural population—if an important electoral constituency—can push politicians to initiate and maintain welfare policies such as the drought relief program in Botswana and the pension schemes in Mauritius. This finding partly diverges from the Western welfare regime literature, where there has been a tendency to focus almost entirely on the urban industrial class and their trade unions. Esping-Andersen (1990, pp. 29-33) did emphasize that small farm owners were critical to the promotion of broad-based welfare policies; and this article demonstrates how the small farmers, although mostly neglected by Haggard and Kaufman, as well as an urban middle class are also critical for welfare policy development in developing countries. Mauritius offers a clear example of a country in which lower- and middle-income classes alike have become committed to broad-based welfare policies, whereas the welfare regime in Botswana is increasingly irrelevant to the emerging middle class.
The role of politics in policy making is further substantiated when emphasizing both the spending and financing side of welfare policies. Although the financing of welfare policies has often been neglected in the welfare regime literature, it is evident from this historical comparison that “free money” in resource-dependent Botswana has provided the political elite with the means to finance social services and hence please popular demands while leaving redistributive transfers and taxation at a minimum. Alternatively, in Mauritius, it is likely that relatively high taxation was acceptable as all income groups benefited from broad-based welfare policies as well as strategies of economic transformation.
The analysis further demonstrates how the development of welfare policies cannot be understood in isolation from the production regime and the related economic strategies. Production regimes affect the power resources of classes, and the case studies also indicate that welfare policies can form part of the economic strategies, as in Mauritius, or economic growth strategies may take precedence over welfare policies, as in Botswana—particularly in regard to social transfers.
The study equally confirms the resilience and continued existence of welfare policies in both Botswana and Mauritius. Even though welfare policies continue to exist—especially if they have a strong support base—we also saw a shift in the welfare regime in Botswana, which was partly caused by the increasing irrelevance of some welfare policies to politically influential groups, for example, the middle class. Thus, policy change is possible, but it will most likely be incremental and bounded by existing policies and political alignments.
Many developing countries lack strong politically and economically independent middle classes. This may affect the long-term development of welfare regimes as the nascent policies may already generate a split between the (poor) beneficiaries of pro-poor welfare policies and better-off income groups who rely on job-related insurance policies and private services and are likely unwilling to bear welfare policy costs from which they benefit little. Whether and how welfare policies can become tools for social cohesion and broad-based development in the South are questions deserving further scrutiny.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.