State Reform and Inclusion:
Changing Channels and New Actors

The nature of exclusion changes over time. Some social groups, such as racial minorities and women, have faced persistent exclusion or discrimination in education and labor opportunities or have been deprived of the possibility of influencing political decisions in their communities or countries. But analyzing patterns of social inclusion and exclusion only from this point of view would be a mistake, because it leaves out many other forms of exclusion that mainly affect the poor—whether or not they are racial minorities or women—and that have an impact on the way Latin Americans of all economic classes perceive how society, the economy, and politics work.

This chapter analyzes the impact on patterns of inclusion and exclusion of three enormously important phenomena that have affected most Latin American countries in the last three decades: democratization, macroeconomic stabilization, and globalization. Obviously these are not the only phenomena that have influenced the constantly changing patterns of inclusion and exclusion in the region. They have been chosen because they have emerged relatively recently and involve a sufficiently large number of countries, making it possible to deduce their most important effects.

One important consequence of these three phenomena has been that they have altered the way the state functions, as well as many of the channels of political, economic, and social inclusion. To a large extent, democratization, economic stabilization, and participation in international trade and financial flows revealed the crisis into which Latin American states had fallen in the early 1980s. The crisis of the state in Latin America was not only fiscal, but also a crisis of administration and legitimacy. It made a decisive contribution to the collapse of autocratic governments, to the reduction of inflation and control of the immense associated fiscal disorder, to the opening to international trade, and to the exposure to international capital flows following the dismantling of controls on external financing and barriers to foreign direct investment. The crisis also set in motion a large number of state reforms, ranging from the way the three branches of government function and relate to one another to the institutions and policies of economic intervention and provision of social services. This “silent revolution” in the Latin American state (Lora, 2007), which is still under way, has changed and will continue to change the patterns of political, economic, and social inclusion and exclusion.

The three sections that follow this introduction analyze how democratization, macroeconomic stabilization, and globalization have expanded the economic, social, and political possibilities of some groups in Latin America and reduced those of others. The last section focuses on the fiscal expressions of the three phenomena, emphasizing the importance of taxation and public expenditure for patterns of inclusion and exclusion in the region.

DEMOCRATIZATION AND POLITICAL REFORMS

In 1977 only four Latin American countries (out of nineteen presidential countries) could be considered democratic (or semidemocratic), with leaders elected in free (although not necessarily fair or clean) elections. Thanks to the wave of democratization that began in the late 1970s and continued through the 1980s until the mid-1990s, all countries in the region can now be considered democratic or semidemocratic (Smith, 2005, chap. 1). Democratization processes and the political reforms of the last three decades have opened channels of political participation for a variety of social groups and altered the balances of political power, which is now less socially and regionally concentrated, although very permeable to the influences of special interests and economic power.

Working Classes and Indigenous Populations in the Democratization Process

The democratization processes in the region were driven by social pressure from groups with growing economic and political power, mainly the organized working and middle classes. Strikes and demonstrations by the working classes in Peru and Argentina were decisive in ousting military governments in 1977 and 1983, respectively. In Chile a multiparty alliance was formed that defeated President Augusto Pinochet in the historic plebiscite of 1988. In Brazil, the famous metalworkers’ strike of 1980 and the formation of the Workers Party aligned the working class with the business sector to form a common front that eventually led to the removal of the country’s military government in 1985. Latin America’s middle classes gave their support to these antiauthoritarian fronts and broadened demands to include human rights, freedom of social organization, and corruption control. The varied forms of expression and organization of “civil society” were crucial in replacing armed confrontation with electoral competition in El Salvador, Guatemala, and Nicaragua in the 1990s and in the Mexican democratic opening in 2000. Consequently, democratization processes essentially sprang from pressure from previously excluded groups under the risk of being left out of political power.[1]

Although the working classes played a decisive role in democratization in the region, in the long run they did not maintain their political influence. Surprisingly, many of the economic reforms contrary to the interests of the organized working class were adopted by governments led by parties traditionally allied with them, such as the Justicialista Party in Argentina, the Liberal Party in Colombia, and Acción Democrática in Venezuela. The liberalization of international trade, privatizations, and certain labor reforms, which were part of a strategy of economic liberalization and participation in the world economy, weakened the power of the working classes and contributed to the general feeling of disillusionment and frustration with traditional political parties.

Democratization also opened spaces for political activism by indigenous populations in some countries.[2] Indigenous movements faced a dilemma of identifying politically with campesino (small farmer) or worker social segments or uniting around their own indigenous identity (see Chapter 10). This is a crucial disjunction that leads, in the first case, to emphasizing social justice for reasons of equality with other groups or, in the second, to emphasizing their differences from the rest of society. Faced with this dilemma, it is not surprising that in countries with relatively small indigenous populations, these groups have mostly taken the first option, whereas the purely indigenous option has been adopted in countries—like Bolivia, Ecuador, and Guatemala—where indigenous populations are relatively sizable (see Box 4.1). But the second option has been a costly conquest, which is still in progress, and not the automatic product of democratization.

Gender Quotas in Legislatures

Between the 1920s and 1960s universal suffrage spread to all of Latin America. In 1929, Ecuador was the first Latin American country to permanently give the vote to women, and Paraguay was the last in 1961 (Smith, 2005: 186). Women’s right to vote was fully recognized in the democratization processes beginning in 1977. However, this has not been sufficient to guarantee the vigorous influence of women in the region’s political life. It was not until the 1990s that measures were taken to increase the political inclusion of women in Latin American countries, specifically guaranteeing their participation in legislatures (see Figure 4.1). About half of Latin American countries adopted gender quota laws in the 1990s, partly for domestic reasons to attract votes among women, as in the case of Argentina in 1991, and partly as a result of international pressure resulting from the Fourth World Conference on Women held in Beijing in 1995, which decided to promote the adoption of quotas to reduce the systematic underrepresentation of women in legislative bodies. As a result, in Argentina, women’s participation in the Chamber of Deputies (lower house) rose from 5 percent to 27 percent, exactly in line with the standard, and in the Senate (upper house) from 3 percent to 36 percent. Costa Rica’s 1997 law was also very effective in increasing women’s participation in the country’s single chamber, from 14 percent to 35 percent (although the quota specified by the conference was 40 percent). For the region as a whole, the adoption of quotas in the 1990s raised women’s participation in legislative bodies by eight percentage points. Quotas are more effective in countries where they are compulsory, where the position of candidates is specified on the electoral list (so that women candidates’ names are not relegated to the end), and where noncompliance (such as loss of the party’s seats in the legislative body) is penalized (see Smith, 2005: 249–53, and Htun, 2004).

Perhaps more significant than the number of women members of congress is participation in cabinet and senior government posts. In 1990 women held only 9 percent of ministerial positions in the region; ten years later their share had risen to 13 percent. In some of the region’s countries—Chile, Colombia, Costa Rica, El Salvador, Honduras, Panama, Venezuela—between a fifth and a quarter of cabinet posts are currently filled by women. In Chile, Colombia, and Mexico, women have held the positions of foreign affairs and defense ministers. And of course, in Argentina, Nicaragua, Panama, and currently in Chile, the presidency has been in women’s hands (although, strictly speaking, only in the last case on their own political merits rather than by extension of their husbands’ power) (Smith, 2005: 249–53).

Patterns of Electoral and Political Inclusion and Exclusion

Regular elections are the key component of democracy, although they do not by themselves guarantee its proper working. Effective exercise of the right to vote in Latin America is far from being total, although it is very high in presidential elections, in which over 70 percent of registered voters usually participate in most countries. Uruguay, Peru, Bolivia, Chile, and Brazil are, in that order, the countries with highest voter turnout, whereas Colombia and Guatemala are at the opposite extreme (see Figure 4.2). Empirical evidence shows that the first elections to establish (or re-establish) democracy in a particular country generate an electoral enthusiasm that later tends to fade. Compulsory voting increases turnout by ten percentage points, although it is difficult to define the impact because its effectiveness depends very much on the sanctions attached to the requirement. Moreover, the percentage varies from one election to another. In general, presidential elections attract more voters than parliamentary or local elections. Turnout is also higher when election campaigns are hard-fought or when elections are dominated by a single issue. It does not seem to be linked to institutional factors (civil rights, political stability, development of parties, political fragmentation), demographic factors (age distribution of the population), or education levels (literacy rates or percentage of the population that has completed secondary studies). In short, turnout patterns in elections remain largely unexplained. It is probable that they are closely related to cultural or historical factors that influence patterns of economic and social inclusion. In this respect, turnout does not seem to be very different from interpersonal trust or any other form of social capital (IDB, 2000, chap. 4).

Since voter turnout in the region is far from total, the question is: Do voter turnout patterns differ systematically from one social group to another? If they do, this will determine whether the groups’ interests are taken into account by the political system. However, the fact that channels of political influence are not limited to the act of voting also has to be considered, since casting a vote is only one form of political participation. Other forms are also important: citizens who are better informed or have more direct contact with parties and candidates are more likely to influence decisions. It could be that turnout and other forms of political participation mutually reinforce each other, which means that the groups that are not heard or have no channels of influence in the political system are the same as those excluded from the act of voting, forming a vicious circle of political exclusion that is difficult to break.

Patterns of political participation by education and economic group were studied in a previous edition of this report based on the Latinobarometer opinion polls in seventeen countries in the region (IDB, 2000, chap. 4). It is worth summarizing the report’s conclusions in regard to these patterns. By education level, differences are small for turnout but large for other forms of participation, such as watching and listening to political news, talking about politics with friends, trying to influence other people’s political opinions, or working for a political candidate. For example, people with at least some college-level education are twice as likely to talk about politics as people who have had only primary school, but the probability of voting among the former group is only slightly higher than among the latter. The trend is somewhat similar for patterns of political participation by income group. There are no discernible differences in voter turnout according to income, but there are for other forms of participation, which tend to increase with income levels. The IDB study also analyzed participation patterns by age group. Surprisingly, participation in the region varies only slightly from one age group to another, although it is found to be slightly lower in younger groups.

It is important to note the relationship between voter turnout patterns and the weakness of political parties. If parties do not represent interests that unite broad sections of the population, then the link between turnout and political inclusion is weak.[3] In countries where political parties are programmatic—such as Chile, El Salvador, Nicaragua, and Uruguay—the population can expect to feel included as a result of the act of voting for the party that most represents their political interests. In contrast, in countries where parties are more prone to clientelism—such as Bolivia, Colombia, Guatemala, Peru, and Venezuela—voting is not sufficient to achieve broad political inclusion, because close links have to be maintained with politicians to obtain the benefits of participation (IDB, 2005: 34).

Patterns of political participation in Latin America do not differ substantially from those in other regions of the world, except perhaps that political participation is distributed more uniformly in Latin American countries than in the United States. In European countries participation patterns are practically identical to those in Latin America. Consequently, based on differences in voter turnout or the other forms of political participation mentioned, it cannot be said that there are very pronounced biases in favor of certain social or economic groups and against others. However, no definite conclusions should be drawn from this analysis because of the difficulty in determining the relative effectiveness of the various forms of participation and because perhaps the most important channels of political influence in Latin America are not captured by opinion polls. In particular, small but economically powerful groups with well-defined interests can have a disproportionate influence on political decisions in the region, since they can organize more easily than larger groups with less economic power and more dispersed interests.

Whether these powerful interest groups can effectively displace other less-organized social groups in a country depends on whether the country’s political system creates incentives for politicians to respond to regional, sectoral, or class interests. For example, soft regulations on election campaign finances facilitate the influence of powerful groups. The same can be said of electoral institutions that oblige politicians to overrespond to geographical interests, because this opens the way for the influence of geographically very concentrated economic sectors. Likewise, some voting systems create incentives for politicians to cultivate personal groups of supporters instead of following their party directives, which presumably represent broader national interests.

According to an index that quantifies the incentives created by voting systems for politicians to cultivate relations with voters or with their party leaders, in Latin America politicians have more incentive to stay on good terms with party leaders than in any other part of the world (IDB, 2000: 193–95). As a result, regional interests generally have little influence on the working of Latin American political systems. Political systems are more susceptible to influences from interest groups that organize nationally in order to reach party decision-making centers directly. Consequently, powerful economic groups with very defined and concentrated national interests can exclude the regional, sectoral, or social interests of groups that do not have national organizing capacity. However, this conclusion has to be balanced with at least three considerations.

First, the electoral systems of some countries do create incentives for cultivating personal relations with the electorate rather than with party leaders. The best-known case is Brazil, where candidates do not have to be nominated by their parties to run election campaigns using the party name and voters can express their preferences for individual candidates on the proposed lists. Second, the election systems of some countries have undergone major reforms. The general trend has been to exchange closed lists for open lists in which voters may choose among candidates of the same party, which increases incentives for politicians to respond more to voters and less to party hierarchies (Payne and Perusia, 2007). Third, the most important change in the channels of influence of regional interests has been political decentralization.

Political Decentralization

Previously, Latin American central governments named officials at lower levels of government and controlled a significantly higher proportion of total public expenditure (Wiesner, 2003). Under this system, local authorities had more incentives to respond to the demands of the central government than to the preferences of the people of their locality. The central government was the body that made most decisions on expenditure, even in local cases. This arrangement reinforced the bias in favor of agents that had national capacity to organize and exert pressure. Public sector labor unions, particularly in the areas of health and education, are a clear example (IDB, 2005; Daughters and Harper, 2007).

The deepening of decentralization in Latin America in the 1990s had two special features that changed the patterns of inclusion and exclusion associated with decentralization. First, the rate of decentralization was not equal among all public expenditure items. The services decentralized between 1996 and 2004 were largely nutrition, public hospitals, maintenance of interurban highways, urban transport services, and regional universities. The second significant characteristic of the recent decentralization process in the countries of the region (with the exception of Brazil) is the local taxation lag. Local authorities now have a series of expenditure responsibilities, but their capacity to levy taxes is very limited, because of very small local tax bases or legal restrictions on designing an independent tax policy (Daughters and Harper, 2007).

Both transformations have had a dual effect on inclusion of the population. Although it is true that election of local authorities and devolution of certain expenditure decisions have involved previously excluded segments of the population, it is also true that this process has been limited because of the low level of local tax collection in the region. In other words, inclusion of the population in local public policymaking in Latin America has been partial: people can now influence the spending decisions of mayors and governors, but there is little concern about the sources of revenue needed to cover such expenditure.

An additional effect of political decentralization has been the strengthened role of nongovernmental organizations and community organizations. (Angell, Lowden, and Thorp, 2001, document this phenomenon very well for Colombia and Chile.) Under centralist schemes, local governments had no need to develop independent administrative capacity; with political decentralization the need exists, particularly on issues related to institutional strengthening and community development. The result is that nongovernmental organizations and community organizations are now participating in the execution and design of local public policies.

STABILIZATION AND MACROECONOMIC POLICIES

Inflation and Economic Populism

In the 1970s and 1980s, Latin America was characterized by persistently high inflation and quite a few cases of hyperinflation (conventionally defined as monthly inflation rates over 50 percent). Populism was the characteristic of the macroeconomic policies of a number of administrations, such as those of Juan D. Perón in Argentina between 1973 and 1974, Salvador Allende in Chile between 1970 and 1973, and Alan García in Peru between 1985 and 1990 (Kaufmann and Stallings, 1991). The economic policies typical of populism produced very marked patterns of exclusion and inclusion of the population, as described below.

In general, populist governments in Latin America during this time assumed power with strong electoral support from the middle classes of formal urban workers in both public and private sectors. These governments argued that the low level of economic activity was a problem of repressed demand. Raising wages was expected to create a virtuous circle of high demand, higher production, and higher wages. For this reason, the starting point of populist measures was significant wage hikes for public and private sector employees (Cardoso and Helwege, 1991). To contain inflationary pressures, governments froze prices and, in some cases, fixed the exchange rate. In the short term, these measures benefited the urban middle classes at the cost of other segments of the population. Higher real wages in the cities—a consequence of nominal wage increases, frozen prices, and appreciation of the real exchange rate—had a counterpart in their impact on agricultural producers, both suppliers of the domestic markets and exporters (Kaufmann and Stallings, 1991). Up to this point it can be said that populist governments kept their political promises in the short term: what they intended and achieved was to favor the social class that supported them politically.

However, monetization of fiscal deficits and the depletion of international reserves revealed the unsustainable nature of these policies. Inflationary pressures put an end to price controls and reversed the trend in real urban wages. The initial economic boom fueled by these measures was followed by deep economic crises, which rapidly wiped out the increase in real income. For example, the real wage of industrial workers in Chile rose 20 percent in real terms between 1970 and 1971, then plunged 11 percent in 1972 and 38 percent in 1973. Real wages in the manufacturing sector did not return to their 1970 level until 1981 (Larraín and Meller, 1991).

The pattern of inclusion and exclusion induced by populist measures responds to criteria of political economy. The poorest population segment in Latin America is concentrated in rural areas, but this group did not have sufficient voice or organization to exercise political pressure and was excluded from public policy decisions. In contrast, the middle class, characterized as urban with formal employment in the public and private sectors, had the organization and voice required to demand measures in its favor, although these measures had only short-term effects (Cardoso and Helwege, 1991).

Another way of interpreting the segmentation produced by economic populism is as a political struggle to determine the social distribution of the inflation tax (which results from the loss of purchasing power of the currency held by the public, as a result of excessive increases in money supply). This tax was very important in Latin America in the 1970s and 1980s: on average 4.4 percent of gross domestic product (GDP) between 1973 and 1983 and 10.8 percent between 1983 and 1987 (Edwards and Tabellini, 1991). In a situation in which a country’s monetary authority can grant unlimited credit at no cost to the central government, how the inflation tax is distributed is decided by those in power. The costs are possibly shared regressively, because the poorest segments of the population and small businesses have no access to financial instruments to protect them from this tax.

These populist experiments financed by inflation taxes seem to be a thing of the past. At the end of the 1980s, eleven Latin American countries had inflation rates over 20 percent, and four suffered price increases of more than 1,000 percent annually. In contrast, over the 2001–2005 period, no Latin American country had an average inflation rate over 20 percent, and only five reported inflation above that figure in any one of these years (with a 41 percent maximum in Argentina in 2002).

A decisive factor in the macroeconomic stabilization processes in Latin America was the granting of independence to central banks to make their own decisions on monetary policy, with the predominant objective (and in some countries the only one) of reducing inflation. Between 1988 and 1996, the central banks of twelve Latin American countries were reformed by law or constitutionally, being granted more independence in the design and conduct of monetary policy as a guarantee of price stability. In this process, governments gave up their discretionary power to finance themselves directly from the central bank, which eliminated the main cause of the inflation tax. As will be shown in the following section, the independence of countries’ monetary authorities also had important effects on the patterns of inclusion and exclusion associated with credit and exchange rate policy.

The End of Directed Credit

Prior to the macroeconomic stabilization of the 1990s, the central governments of Latin America had considerable ability to direct credit to influential economic sectors or groups. This was feasible because public banks were very important within the financial system, because the central bank was an important intermediation agent, or because the government strictly regulated allocation of credit by private banks. In the late 1980s, a significant percentage of total lending in Latin America was directed by the government using one of these three mechanisms: around 30 percent in Colombia, 35 percent in Mexico, and 40 percent in Argentina, and up to 80 percent in Brazil (Morris et al., 1990). In these financial systems in which the government played a central role, access to credit was determined by the ability of economic agents to exert political pressure. Agents that were unable to exert pressure were excluded and had to pay high financing costs (Edwards, 1995).

Brazil represented the region’s most extreme case of discretionary directed credit. First, public banks were and still are central players in the country’s financial system: in 2002, 43 percent of the total assets of the Brazilian financial system were held by public banks (Galindo, Micco, and Panizza, 2007). Moreover, in the 1970s and 1980s, national development banks, particularly the Brazilian National Development Bank (Banco Nacional de Desenvolvimento Econômico e Social, or BNDES), allocated a large amount of funds to the industrial sectors, which were protected at that time. In general, credit was granted to the most powerful sectors in the economy, especially heavy manufacturing, finance, and large-scale agriculture (Frieden, 1991). Under the argument that these sectors were fundamental for national development, these policies excluded highly profitable projects in other sectors.

In recent years, the trend toward directed credit in the region has been reversed, and market mechanisms have been given a greater role. Privatization of public banks, the independence granted to central banks, and deregulation of private lending have cut off old, vocal industrial groups from cheap or subsidized credit. At the same time, new credit options have emerged, such as microcredit or lines of credit for small- and medium-sized enterprises, including sectors previously excluded from the financial system (IDB, 2004). Another recent phenomenon in Latin America has been the development of local bond markets. Here two opposing forces have come together, because there is evidence that growth of public debt bonds has stimulated bonds of private origin. However, an excessive level of domestic public debt is simultaneously crowding out the borrowing capacity of the private sector (IDB, 2006b, chap. 7), which excludes formal companies from this nonintermediated financing mechanism.

Exchange Rate Policy

Exchange rate policy has undergone a transformation similar to that of credit policy, and the patterns of exclusion and inclusion that it generates have also changed. To understand this transformation, it should be borne in mind that a country’s exchange rate system affects each population group differently. Regimes with fixed or fairly inflexible exchange rates can be expected to produce low inflation rates and real appreciation, which benefits the urban middle and upper classes by increasing their purchasing power. This scenario, however, harms agro-exporters and other producers of tradable goods. A flexible exchange rate regime, accompanied by moderate inflation and real appreciation, produces the opposite effect (Blomberg, Frieden, and Stein, 2005).

Some empirical works have found that trade liberalization has been a breaking point in the relation between exchange policy and the interests of different sectors in Latin America (Frieden, Ghezzi, and Stein, 2000). In the previous context of high tariffs, the protected economic sectors had no interest in pressuring for an exchange rate policy in their favor. After losing their tariff protection as a result of trade liberalization, the sectors were interested in compensating for the loss by pressuring for more flexible exchange rate policies with a trend toward real depreciation. Frieden, Ghezzi, and Stein (2000) found a positive relation between the weight of manufacturing industry in output and the probability that a country had a flexible exchange rate system. This relation has strengthened in recent years as barriers to international trade have been removed.

In conclusion, before the effects of globalization, the objective of exchange rate policy in Latin America was to protect the real income of the urban middle and upper classes, without taking into account the economic sectors protected by trade policy. As trade protection declined, these sectors’ power to determine exchange rate policy increased, and the interests of the urban middle and upper classes lost ground.

Transformation of Industrial Policy

Public ownership of companies and privatization processes are one aspect of relations between the state and the productive base; another is industrial policy, which has changed drastically, particularly as a public policy response to the globalization process. The transformation of industrial policy in Latin America and the impact of the change on social exclusion and inclusion are briefly described below:

Industrial policy prior to the globalization process was very uniform among the countries of the region in the framework of import substitution policy (Melo, 2001). As mentioned earlier, its main instruments were tariff protection, directed credit, direct subsidies, and exchange rate controls. Its center of attention was the manufacturing sector, especially heavy industry, and its objective was to develop previously nonexistent economic activities to supply domestic markets.

Under the old industrial policy, the main beneficiaries were business owners and formal employees linked to the protected sectors. Urban in origin, both segments of the population succeeded in protecting companies and formal jobs at the cost of high domestic prices, valuable fiscal resources, and a bias against the agricultural and export sectors (Melo and Rodríguez-Clare, 2007).

The transformation of this industrial policy into that which exists today in the countries of the region was not a gradual or linear process. Structural adjustment policies in the late 1980s and early 1990s led to temporary abandonment of the type of intervention represented by the old industrial policy in most Latin American countries. Later, in the mid-1990s, industrial policy in the region began a slow recovery that is still evident (Peres, 2005). However, unlike those under the import substitution model, the new industrial policies are characterized by (a) being much more heterogeneous across countries and (b) having the clear purpose of stimulating potential export sectors. In other words, there has been no return to the previous industrial policy model (Machinea and Vera, 2006; Ramos, 1996).

Melo (2001) and Peres (2005) offer a typology of the new spectrum of industrial policies in the region.[4] A group of countries, among which Brazil is a good example, have focused on supporting the economic activities in which they have clear dynamic comparative advantages (for example, biotechnology) or have technological externalities that could be very valuable for the national productive base (for example, information technology in Costa Rica). In both cases, the ultimate objective is not to supply the domestic market, but to develop new exports. Another group of countries, of which Mexico and Colombia are good examples, support existing clusters of productive activities. The objective is to raise competitiveness and embrace the international competition that these activities face without granting direct subsidies to the companies involved. A third and last group, of which Chile is the outstanding example, has applied horizontal or neutral industrial policies across sectors. The objective is to stimulate the productivity of the largest number of economic activities without interfering in factor markets for allocation of resources.

The diversity of current industrial policymaking in the region is reflected in the resulting change in patterns of exclusion and inclusion. Neutral policies and policies to promote sunrise sectors, unlike cluster policies, are less likely to be captured by business groups and can be more equitable in sectors that did not previously receive public support. Nonetheless, cluster promotion policies are much more inclusive than the old industrial policies because they are targeted at sectors other than manufacturing and because they are not based on protectionist measures such as tariffs or market quotas. Even so, current industrial policies, particularly those targeted at clusters, have very few elements that promote investment in human capital in the sectors that receive support. It could be said, then, that modernization of the labor force is the great challenge for modern industrial policy in Latin America.

Although it is true that Latin America’s new industrial policies are more inclusive than the ones they replaced, it is also true that old patterns of exclusion persist as a result of the concentration of economic power. Such is the case of the business class in Guatemala, which, thanks to its voice and cohesion, has succeeded in retaining generous tax exemptions, another way of capturing industrial policy (IDB, 2005).

Relations between the state and companies in Latin America are reflected in the trend in informal employment in the region.[5] There is evidence of a secular increase in this type of employment in the region in recent decades, which demonstrates the state’s inability to create the necessary incentives for formal companies to assimilate the growing supply of workers, either under the old protectionist model of industrial policy or under the current model in any of its variants. According to Thorp (1998), in 1950 only 8.7 percent of urban manpower in Latin America was linked to informal employment. At that time, companies and formal employment were covered by protectionist policies, and demographic pressure stemming from urbanization had not materialized. Two decades later, this proportion had increased to 11.5 percent, threatening employment security in the region.

The World Commission on the Social Dimension of Globalization (2004) documents the rapid growth of informal employment in the region in the 1980s and 1990s. According to the commission, the percentage of self-employed workers in Latin America in the 1980s in activities other than agriculture was 29 percent, only three percentage points above the world average.[6] In the next decade, this percentage increased to 44 percent, twelve percentage points above the world average and very close to the average in Africa (48 percent, according to the information in Figure 4.3). Protected urban companies did not succeed in growing sufficiently to absorb the flow of migrants from rural to urban areas. Combined with the high costs associated with formal employment, this factor has turned informal and badly paid jobs into a persistent phenomenon.

According to figures from the Economic Commission for Latin America and the Caribbean (ECLAC, 2006b) in its latest report on the social panorama of the region, this trend has reversed itself slightly in some countries in this decade (especially Argentina, Chile, and Costa Rica) but has continued in others (Colombia, Nicaragua, Paraguay, and Venezuela being the most extreme cases). Even more worrisome is the fact that the income of people in this class of activities plummeted during the last decade in almost all countries of the region, with only Chile, Ecuador, El Salvador, and Panama excepted.[7] These problems are studied in more detail in Chapter 5.

GLOBALIZATION AND TRADE LIBERALIZATION POLICIES

In the 1990s, the increased participation of Latin American eco­nomies in international trade, finance, and technology flows and, to a lesser extent, international migration flows affected wages, employment, and social security conditions and exposed workers and their families to more opportunities, as well as to higher risks, than in the past. As happened with the phenomena of democratization and stabilization, globalization altered patterns of economic and social inclusion and exclusion for the benefit of some groups, but to the detriment of others.

Trade Liberalization and Its Effects on Labor Exclusion and Inclusion [8]

To benefit from the possibilities of growing international trade and obtain the blessing of investors and international creditors once again interested in Latin America after resolution of the region’s debt crisis, in the mid-1980s Latin American governments began to put in place a set of reforms to open their economies to trade, finance, and investment. [9] The central element of this external liberalization was reduction of import barriers, which protected local production, especially in the industrial sectors. Between the mid-1980s and early 1990s, all of the region’s countries began trade liberalization programs with reductions of at least fifteen points in the average tariff rate. Tariffs dropped from average levels of 48.9 percent in the years prior to the reform to 10.7 percent in 1999. Nontariff restrictions, which applied to 37.6 percent of imports prior to the reform, affected only 6.3 percent by the mid-1990s.[10] But the lower tariff and nontariff restrictions fueled imports, which grew as a proportion of GDP in most countries. For the region overall, import penetration grew from 22.6 percent in the 1983–1985 period to 36.2 percent in the 1998–2000 period. Export indicators also rose, although much less steeply, from 23.3 percent to 29.6 percent of GDP between the two periods.

Although in popular opinion, trade liberalization and overall unemployment are closely linked phenomena, this belief lacks empirical support. Unemployment rates in the region did not increase following the liberalization measures or as a result of the increased entry of imported products into economies. Likewise there is no basis for stating that total employment fell. However, this does not mean that liberalization did not have destructive effects on jobs in specific sectors, as in fact it did. Surprisingly, however, these destructive effects were relatively modest. Changes in the sectoral composition of employment were very small in comparison with the size of the measures and what all studies preceding the enactment of the liberalization measures had predicted.

Since reallocation of employment between sectors as a result of trade liberalization was very low, companies could be expected to adjust to liberalization in some other way. In part they did so by improving their efficiency and redirecting their production into more profitable activities. But most of the adjustment fell on workers through lower wages. In the case of Mexico, in companies affected by a forty-point tariff cut, real wages fell by an estimated 8–10 percent. For the overall manufacturing sector, tariff cuts at the end of the 1980s produced an estimated 3–4 percent fall in wages. Elimination of quantitative controls on imports may have had an even greater effect that is difficult to quantify (Revenga, 1997). In Colombia, where the average tariff dropped from 50 percent in 1984 to 13 percent in 1998, the effect on the average wage in the manufacturing industry was also 3–4 percent[11] (for the initially more protected industrial sectors, it was up to 7 percent).

It is surprising that the effects of trade liberalization on wages have been relatively severe in comparison with modest changes in employment and its composition. The most probable explanation for this phenomenon is that workers shared the rents (and inefficiencies) that protection gave to companies. Tariff reduction could be accommodated without great changes in employment by improving productivity and eliminating these rents. The disappearance of protection rents also weakened the power of unions, which lost influence in wage negotiations and in maintaining or expanding industrial employment. Not surprisingly, in most countries of the region, union participation rates fell in the 1990s to only 16.3 percent of the labor force in 1991–1995 from 20.1 percent a decade earlier or 25 percent in the second half of the 1970s.[12] However, there were other phenomena that contributed to the loss of union influence, especially reduction of employment in the public sector, expansion of temporary employment, and changes in legislation that governs the operation of unions.

Traditional forms of hiring in the region have been partly displaced by new arrangements, such as outsourcing of services and temporary employment.[13] However, increased international trade has played at most a marginal role in this process, which on the demand side of labor has been driven by technological and organizational change, and on the supply side by the demand for greater flexibility, especially for women workers.[14] Evidence suggests that increased international trade has produced more informality only in countries with more rigid labor regulations.

Although competition from imported products has reduced wages in the affected sectors (and in some countries may have contributed to increasing informality), the emergence of new export sectors has created new labor opportunities in Latin America. Contrary to fears on this issue widely covered by the international press, studies conclude that workers in new agricultural export sectors or in maquiladora companies receive higher pay and have better working conditions than workers in any alternative available job. These conclusions are confirmed by the workers themselves in export zones of nontraditional crops in Guatemala or in the maquiladoras of Chihuahua and Ciudad Juárez in Mexico, to mention only two examples.[15]

Technological Change and Wage Gaps

While increased trade penetration has tended to depress workers’ wages in the affected sectors, technological change, which has accelerated with globalization, has benefited Latin America’s more-skilled workers. The widening wage gap between skilled and unskilled workers has been one of the phenomena that has generated the most reaction against globalization in the region.

Although the widening of wage gaps between workers according to education level has been an important phenomenon in Latin America, it has been less pronounced than is often claimed. A comparison of the wage income of workers who had completed tertiary studies with those who had completed secondary studies reveals an 18 percent increase (average for twelve countries in the region) in the income gap between the two groups during the 1990s: the ratio of the income of workers with tertiary education to that of those with secondary education increased in Latin America from 2.3 in the early 1990s to 2.8 in the early part of this decade.[16] If the comparison is between workers with completed tertiary studies and workers with completed primary studies, the increase in the gap is 7 percent. In contrast, if the calculation is between workers with completed secondary and completed primary education, the gaps narrowed slightly during the decade.

These trends are not common to all countries. For example, considering the gaps between incomes of workers with tertiary and secondary education, Argentina and Nicaragua show important increases (53 percent and 24 percent, respectively), whereas Brazil, Honduras, and Panama show modest reductions. In several countries the trend toward expanding gaps at the start of the trade liberalization process has halted or reversed itself in recent years. In Mexico, the trend ended in 1994, when the North American Free Trade Agreement (NAFTA) entered into force, and in Colombia the gap worsened severely in the early part of the 1990s but reversed completely in later years.

Numerous studies have analyzed the causes of widening wage gaps in Latin America. Although several have found some relation with trade liberalization processes, most tend to conclude that the trend is due to technological changes associated with certain types of imports. A study that included Argentina, Brazil, Chile, Colombia, and Mexico found that wage gaps by manufacturing subsector had common patterns for technological reasons.[17] In synthesis, technological change seems to be a much more important cause of wage inequality than international trade.

The gender wage gap has followed a different path than gaps by education level (Table 4.1). Some studies show that the gap in remuneration for work between men and women with similar skill levels has narrowed or has remained stable for the last twenty years (see Tenjo Galarza, Ribero Medina, and Bernat Díaz, 2004, for the cases of Argentina, Brazil, Colombia, Honduras, and Uruguay; Ñopo, 2006, for the case of Chile). Even so, gender differences in wages persist (see Chapter 2).

FISCAL INCLUSION AND EXCLUSION

The previous sections of this chapter showed how democratization, stabilization, and globalization changed the patterns of inclusion and exclusion of social groups in Latin America. Democratization opened spaces for political participation by the working classes, indigenous groups, and women, but also strengthened the influence of interest groups with ability to organize at the national level or exert direct influence on political parties. Stabilization limited the access to cheap credit enjoyed by certain sectors and privileged groups, but also reduced the influence of the urban middle classes, whom price controls, artificial fixing of exchange rates, and increases in minimum wages of earlier periods had tried to benefit, albeit only in the short term. Trade, financial, and technological globalization helped erode the power of the urban middle classes, especially low-skilled workers, while strengthening the influence of workers with higher levels of education and owners of capital.

Democratization, stabilization, and globalization also changed the way a country’s social and economic groups relate to the country’s tax administration and public expenditures. This is no surprise, since fiscal policy is the way the state distributes resources to all sectors of society for collective purposes and for the needs and interests of specific groups. Consequently, participation by all the different social and economic groups in collection and allocation of fiscal revenue is a decisive factor in patterns of inclusion and exclusion.

The most important influence of democratization on fiscal systems should have been the adoption of strongly redistributive tax systems, as the median voter theory predicts. This theory predicts that if everyone voted, taxes and the size of governments would be greater in more unequal societies. Because of the income concentration effect, in Latin America there is in fact a tendency to impose high tax rates on individuals, with higher rates in countries with more inequality, also consistent with the median voter theory. There is also evidence that rates of income tax collection are higher in countries with greater inequality, even after the effects of differences in income per capita and voter turnout are isolated (Lora, 2006). However, collection rates in the region are very low by world standards and increased very little with democratization, possibly because of the influence of high-income groups on the design and application of tax rules. As is very well summarized by Richard Bird (2003: 41), a recognized expert on Latin American tax systems, the rich in Latin America have many ways to avoid paying direct taxes: “First, they can block progressive legislation; they can introduce incentives and exceptions to dilute its effects (always with the argument of ‘national interest’); they can corrupt the tax administration or use their resources to tone down or delay its applications in legal ways; or they can escape with their funds from the jurisdiction.” Also, as will be discussed later in this chapter, financial globalization has weakened the treasury’s capacity to tax internationally mobile capital, and trade globalization has eroded taxes based on international trade. Governments have been forced to shift to higher indirect taxes, particularly value-added taxes (VATs), which is not an expected consequence of democratization.

Regarding public expenditure, democratization seems to have had a deeper effect, consistent with raising the share of social spending in total expenditure. At the world level, democratization processes have led, with a lag of some years, to an increase in social spending of 3 percent of GDP (Baqir, 2002). This phenomenon has also occurred in Latin America. Average public social spending in the region, as a percentage of GDP, grew around three percentage points between 1990 and 2003, from 9.6 percent to 12.8 percent, and real levels per capita increased substantially. In the early 1990s, the countries of the region allocated an average of US$314 per person (in constant 2000 dollars)[18] for items of public social spending. Thirteen years later, average spending per capita was 45 percent higher: US$457 per capita (ECLAC, 2006a).

However, the influence of democratization on the distribution of the benefits of this social spending is less clear.[19] In Chile, where social spending is currently very progressive, democratization has been a decisive factor (see Box 4.2). However, in the rest of Latin America, there is no evidence that democratization or later political reforms have altered the distributive patterns of social spending, which tend to be approximately equal by income level.

With stabilization, the monetary sources of financing enjoyed by Latin American treasuries were closed, reducing the discretionary power of central banks (and through them central governments) to allocate resources, as shown earlier, but also imposing discipline on spending demands by legislatures. Adoption of fiscal responsibility laws in several countries,[20] which put caps on expenditure, the fiscal deficit, or both, also strengthened fiscal discipline.

When the sources of monetary financing disappeared, alternative ways to cover expenditure had to be found. Financial globalization, which accelerated with the Brady Plan in 1989 (conversion of long-term government debt instruments after the crisis of the 1980s), opened the possibility of external financing of fiscal deficits, at least temporarily. At the same time, financial globalization limited the possibilities of taxing capital income because of increased capital mobility and the need to compete for foreign direct investment. Consequently, the position of large national and international owners of capital was strengthened against other sectors in taxation decisions. For its part, trade globalization imposed very narrow limits on taxes on international trade, which in many countries were an important source of fiscal revenue. To compensate for these adverse trends and respond to growing pressure for spending generated by democratization and other factors,[21] governments had to turn to other sources of tax revenue, particularly VATs, and maintain or increase taxes on and payroll contributions by formal workers. In some countries with relatively mature social security systems, such as Colombia, workers’ contributions to the system were raised considerably to cover the large deficits that resulted from the need to pay generous pensions to growing numbers of pensioners whose contributions had been insufficient or badly managed.[22]

Who Pays Taxes and Who Does Not?

The changes in the tax structures in Latin America as a result of the forces of democratization, stabilization, and globalization primarily involved composition rather than size. On average, tax receipts (not including social security contributions) totaled 16.3 percent of GDP in 2003, practically unchanged since the mid-1980s (15.4 percent) despite numerous reforms. The tax burden in Latin America is 6.8 percentage points of GDP lower than world standards. Brazil and Argentina, with tax burdens of 21 percent and 18 percent, respectively, are the two countries with the highest receipts; Guatemala, Panama, and Paraguay, with burdens of approximately 10 percent of GDP, have the lowest tax burdens. The main shift in composition has been in favor of VATs. Whereas receipts from direct taxes in the region have averaged around 4 percent of GDP in the last two decades (and are 3.5 points below the world level), VAT receipts rose from 2.9 percent of GDP at the end of the 1980s to 5 percent in 2003, offsetting the decline in taxes on international trade and numerous minor taxes. Total indirect taxes collected in Latin America (7.9 percent of GDP) do not differ from the world average.

As a consequence of these changes in composition, Latin America’s lower and middle classes now experience the impact of the tax burden more directly. In the past, the effect of taxes (and other forms of protection) on imports strongly affected the purchasing power of these segments in relation to consumer goods, but not directly as in the case of VATs. For this reason, and because of the widely held opinion that the VAT is by nature a regressive tax, the middle and lower classes in the region have begun to have a politically important weight in the tax debate which they did not have in the past. In response, congresses have generally preferred to exclude the main items in popular consumption baskets from the VAT base or to tax them at lower rates. This is reflected in the high tax expenditure of indirect taxes in Colombia, Ecuador, Guatemala, and Mexico, where lost revenue resulting from VAT noncollection and specific tax exemptions exceeds 25 percent of potential tax collections (see Table 4.2). The resistance of the middle and lower classes to VATs has led to a constant search for other sources of tax revenue or to a cap on government spending. It may also have been an important factor in establishing spending programs targeted at the poor, for example, the conditional cash transfer programs, such as Oportunidades in Mexico and Bolsa Familia (Family Grant) in Brazil, that have been adopted in nine countries. What is perhaps even more important: the visibility of VATs may have created incentives for the lower and middle classes to increase their vigilance over the use of public funds and participate more actively in political life. As previously noted, in Latin America there are no pronounced biases in patterns of political participation by social class. There (as in the rest of the world), higher tax burdens tend to be associated with increased political participation.[23]

Those in the working classes with medium or high income from employment in the formal sector have also begun to perceive the fiscal and parafiscal burdens more directly. Through income withholding, wages are relatively easy to tax. This mechanism is applied to workers who earn above the taxable minimum, which on average in Latin America is 230 percent of income per capita. The formal working classes are also taxed with contributions to pension systems, which have increased in most countries. Based on averages from eleven countries, contributions to pension systems increased from 22.7 percent of gross wages in the region before the reforms to 26.6 percent afterwards.[24] In addition to the parafiscal burden imposed by contributions to pension systems, in some countries other payroll taxes are imposed to fund transfers to training institutes and other social service programs. The case of Colombia, where the total parafiscal burden is more than 50 percent of wages, is exceptional among countries of the region (see Bernal and Cárdenas, 2003).

However, as discussed in Chapter 5, growing fiscal and parafiscal burdens on wages have contributed to the informalization of employment in the region. Informal workers are by definition excluded from fiscal policy with respect to tax and social security benefits. In fact, only 27 percent of the Latin American labor force is affiliated with social security systems, and only one-quarter of those over age 65 receive pensions. Since rates of social security membership are higher in the higher income groups, expenditure on pensions is the most regressive of all types of social spending. The most extreme case is again Colombia, where 80 percent of pensions are paid to the richest quintile of the population. The most equal pension spending among the countries of the region is in Costa Rica, which has an effective system of family-wide social security membership with broad coverage.[25]

The coverage patterns in Latin America show two different social inequality mechanisms that complement one another. The first one is social security benefits concentrated among the richest. This goes hand in hand with the exclusion of poor people from both contributions and benefits of the social security system. The second is generational segmentation of the population, in which older generations create tax pressures on younger ones and there is no guarantee that younger generations will receive the same social security benefits as the previous ones have.[26] Both exclusion processes are illustrated in the inconclusive fiscal comedy depicted in Box 4.3.

There has been little success in incorporating owners of capital into payment of taxes, as a result of globalization and the influence that companies and high-income individuals have on decision making and implementation of tax policy. The Economic Commission for Latin America and the Caribbean (ECLAC) estimates that foregone tax revenues (tax expenditures) due to exemptions from direct taxes represent more than 40 percent of potential receipts in Chile, Colombia, Ecuador, Guatemala, and Mexico (see Table 4.2).[27] The sectors that benefit most from these exemptions are usually those with high percentages of exporters, especially in the maquiladora sectors, and companies with foreign investment. However, a wide range of sectors enjoy exemptions, depending on the country.[28] Moreover, most studies conclude that tax breaks and exemptions play a very modest role in corporate investment decisions and are not the best use of fiscal resources (see Cetrángolo and Gómez Sabaini, 2006; Hernández et al., 2000).

Who Benefits from Public Spending and Who Is Excluded?

It is not easy to determine who benefits from expenditure in areas such as security, environment, justice, or investment in infrastructure, because these services are nonexclusive public goods, in the sense that the benefit that some individuals receive does not exclude others from benefiting as well. In other cases, there may be exclusive benefits (for example, use of road infrastructure), however, there is no available information that relates spending to beneficiaries.

Most analyses of the impact of public expenditure concentrate on the social services of education, health, and social security and on some subsidies for utilities.[29] On public social spending, the most general conclusion that can be extracted is that it tends to be distributed more or less equally among income groups, which is to say that social spending is inadequately targeted to the lower classes. With the exception of Chile, where social spending is clearly progressive, total spending on education and health in Latin American countries reaches all large social groups equally (see Figure 4.4). In general, spending on primary education is progressive, but at other levels of education, expenditure tends to be concentrated on higher income groups because of the higher dropout rate in lower income groups. In health, distribution of spending differs greatly from one country to another: it is clearly progressive in Argentina, Chile, Costa Rica, El Salvador, and Honduras, approximately flat in Colombia and Uruguay, and strongly regressive in Bolivia. These distributive patterns were not very different a decade ago, which suggests that higher social spending has not been accompanied by increased participation in social services (IDB, 1998a).

Worse than the distribution of social spending on basic education or health in the region is that of spending on university education and subsidized utilities charges (IDB, 1998a, chap. 8), because consumption of these services is naturally concentrated in the middle and upper classes. But the fact that social spending and subsidies are not concentrated on the poor, or even that they benefit the middle and upper classes rather more than the lower classes (regressive distribution), does not necessarily mean that they aggravate disparities in income distribution. This happens only when the distribution of benefits is more concentrated than income distribution, as in the case of pension expenditure in some countries, for example, Colombia, where 80 percent of pensions benefit the richest quintile (see Figure 4.5).

Analyses of the impact of social spending assume that all benefits of such spending are received by users of social services. But this is not necessarily so. The beneficiaries of social spending, and public spending in general, are also teachers, doctors, nurses, and other public officials who receive wages, which are the most important component of social spending. Public sector payrolls are another case, like that of pensions in some countries, in which expenditure tends to benefit the middle and upper classes substantially more, to the point of increasing income concentration in most Latin American countries.[30] This, of course, largely reflects the fact that public employees in the region have higher than average levels of education. However, for their levels of study, experience, and dedication to work, they are also better paid than their counterparts in the private sector. For example, Panizza (1999) found that in the mid-1990s, the average pay of public employees in eight Latin American countries[31] was 14 percent higher than that of their equivalents in the private sector. Jobs in public service in the region have other features that make them more attractive than other jobs, such as stability, attractive retirement conditions, and social benefits. Consequently, they offer an example of exclusion, because workers who are outside public service have very limited possibilities for access to these benefits. The few statistics available on jobs and pay suggest that public employment in the region may have become even more exclusive in recent decades. According to a study by the Bank’s Dialogue on Transparency and Public Management, public employment in the region overall fell from 5.4 percent of the population in 1995 to slightly over 4 percent in 1999. However, the reductions in employment were not accompanied by a reduction in the value of the payroll: in most countries expenditure on the public payroll actually increased after staff reduction measures (Echebarría and Cortázar, 2007).

In conclusion, fiscal policy in Latin America in the wake of the democratization, stabilization, and globalization processes has favored the inclusion of some groups, especially the lower classes, which have begun to participate in some of the benefits of spending and perceive tax burdens more directly. However, the richest families and owners of capital have maintained or strengthened their tax privileges, while the middle and upper working classes have increasingly split into an included group and another group that is excluded from both sides of the fiscal equation. These processes of fiscal exclusion and inclusion are a truly inconclusive social comedy (Box 4.3). How this comedy is resolved will depend on how Latin American societies solve their serious problems of fiscal exclusion.