Employment

Abstract

This chapter examines the stratification effects of social protection institutions on employment. The conventional approach, supported by current theories of welfare institutional development, is to assume that the structure of employment shapes social protection institutions. This makes sense when researching the origins of welfare institutions. However, once social protection institutions are in place, they exercise considerable influence on employment. The direction of causality is now from social protection to employment. The discussion in the chapter explores the stratification effects of social protection institutions on the structure of employment. The chapter discussed social protection contribution to three changes in employment in the region: tertiarisation, job polarisation and relocation, and participation in global production networks. Approaching the association between social protection and employment as a causal link brings new insights on social protection institutions in the region.

This chapter studies the stratification effects of social protection institutions on employment, the first of the three main causal effects identified.

The literature on social protection and employment has explored their linkages in detail in Latin America. Perhaps the salient characteristic of employment in the region is the large share of workers in informal employment. They include independent workers and dependent workers in small enterprises. Workers in informal employment lack access to pension schemes. In fact, common measures of informality are the residual of measures of participation in pension plans, whether in occupational insurance funds or in individual retirement savings plans. Informality is defined by the selective reach of social protection. There is broad consensus on the close correlation existing between employment status and participation in pension plans.

Consensus breaks down when the issue is to identify the direction of causality in the relationship between social protection institutions and the structure of employment. The conventional view is that the structure of employment places hard restrictions on the reach of pension schemes. Own account workers may have few incentives to participate in pension schemes if their pay flickers around a subsistence level and their employment is precarious. Skilled workers with stable employment and good pay have strong incentives to participate in pension plans offering high and subsidised replacement rates on retirement. It is possible to turn the direction of causality around and make a case that social protection institutions contribute to the stratification of employment. Occupational insurance plans support skilled workers in good jobs. Large firms employing skilled workers might find it advantageous to offer occupational pension schemes for the purposes of enhancing productivity incentives as well as recruitment and retention. Proponents of individual retirement savings plans argued that removing the redistributive features of occupational pensions altogether would encourage job mobility and improve productivity. Levy argues that the overlap between insurance and assistance generates incentives for informality (Levy, 2008; Levy & Cruces, 2021). Arguments in favour of decoupling protection from employment status, as in social assistance or minimum guaranteed income proposals, aim to moderate the employment stratification effects of social protection institutions (Barrientos, 2009, 2012).

The framework developed in Chap. 3 distinguishes the period associated with the emergence of occupational insurance funds from the subsequent periods when social protection institutions are in place. The framework hypothesises that the process of industrialisation and the associated emergence of an urban working class was a causal factor in the eventual expansion of occupational insurance funds. The first incorporation led to the emergence of state-sanctioned occupational insurance funds. The direction of causality runs from changes in the employment structure to the emergence of social protection institutions. Research on social protection since has continued to assume the same direction of causality applies. Research and policy studies of social protection in Latin America often ascribes the perceived limitations of social protection institutions to primordial employment structures, e.g. the persistence of informality, the size of the middle class, reliance on primary production, scarcity of skilled work, or the size structure of firms (Bertranou & Marinakis, 2020; Carnes & Mares, 2015; Eslava et al., 2021; Schneider, 2013). Social protection institutions are the dependent variable.¹

The discussion in this chapter offers an alternative take on the direction of causality following the establishment of social protection institutions. Industrialisation did cause the emergence of social protection institutions in Latin America, and elsewhere. However, when these institutions are in place, they exert observable influence on the employment structure, as well on protection and incorporation. The discussion below focuses on an examination of the causal influence of social protection institutions on employment. As the materials discussed demonstrate, reversing the direction of causality is justified. Social protection institutions are a stratification mechanism. They influence the structure of employment, the pattern of employment relationships, and the productivity of employment. In fact, examining the influence of social protection institutions on employment adds to our understanding of both and provide new insights into the nature of social protection in the region.

The structure of the chapter is as follows. Section 5.1 discusses the direction of causality in more detail. Sections 5.2, 5.3, and 5.4 respectively on Industrialisation, Limited Fordism, and Earnings deals with occupational insurance funds. Sections 5.5, 5.6, and 5.7 respectively on Tertiarisation, the Knowledge economy, and Global production networks discuss the influence of neoliberal and post-neoliberal social protection reforms on changes in employment. A final section concludes.

5.1 Social Protection and Employment: Direction of Causality

Current theories of welfare institutions emphasise the distribution of employment and jobs as a dominant factor explaining the configuration of social protection institutions. This approach is rooted in the view that modern social protection institutions can be traced back to industrialisation. Compared to the diversified livelihoods experienced in rural areas with subsistence agriculture as a fallback, employment in urban industrialised economies is binary. Workers are either employed or unemployed. Unemployment has direct implications for their wellbeing (Atkinson, 1989). The emergence of mutual assistance and public assistance is therefore a consequence of industrialisation. From this perspective, social protection institutions move in lockstep with changes in employment, such that changes in the latter will inevitably bring about change in the former. Social protection policy debates often begin by examining changes to the structure of employment over time before moving on to identify policies appropriate to addressing the outcome arising from employment change (Bertranou & Marinakis, 2020; ILO, 1972).² From this perspective, the structure of employment, and associated socio-economic outcomes, can explain the dynamics of welfare institutions.

In practice, once institutions are in place, their weaknesses or resilience feedbacks into employment, raising the potentially autonomous role of institutions in reinforcing or preventing change. These approaches grant social protection some policy autonomy. A very influential approach can be traced to Polanyi’s (1957) masterly analysis of capitalism. Polanyi argued that capitalism and social protection had distinct and conflicting logics. Capitalism is dominated by a logic of commodification. The forced extension of commodification to non-commodities like labour and land was essential to sustain the accumulation of capital. Social protection institutions emerged to address the resulting fractures: the impact of accumulation on the destruction of the environment and the threats to social reproduction enforced on workers because of a race to the bottom on wages. Polanyi argued that social protection was directed by the logic of solidarity, in contradistinction to the logic of commodification at the core of capital accumulation. Social protection institutions prevent the internal dynamics of capitalism from leading to its destruction. Polanyi described this process eloquently as a double helix, in which one, or the other, logic came to dominate social processes at times. Social protection came to the forefront when capitalism was threatened by crises. Commodification, taken to its limits, results in crises and is rescued by a reassertion of social protection institutions. It is hard not to note the social-democratic overtones in this dualism. Note that in the Polanyi approach, social protection and commodification have independent dynamics, one is called to dominate only when the other has reached its (temporary) limits.

The present work proposes an alternative perspective. It supports the view that industrialisation gives birth to social protection. This is the case in Latin America. Industrialisation resulted in the spread of Bismarckian occupational insurance funds. However, the expansion of social protection institutions in the region does not originate in solidaristic values, independent of the structure of production and employment, as argued by Polanyi. Social protection institutions originate in changes in the employment structure associated with industrialisation. However, once social protection institutions are in place, they stand in a causal relationship with employment. Social protection institutions help regulate the very structure of employment, not just its consequential welfare outcomes.³ They contribute to regulate the scale of available employment, the distribution of skills and productive quality of employment. Social protection institutions have observable effects on employment stratification. Now causality goes from social protection institutions to employment, not the other way around.

5.2 Industrialisation and Social Protection Expansion

The emergence of social protection institutions in Latin America in the middle of the twentieth century is explained by industrialisation. Figure 5.1 shows estimates of the ratio of gross value added in industry to gross value added in agriculture in the ‘long’ century. The data are sourced from the Montevideo-Oxford Latin American Economic History database (MOxLAD, n.d.). The figure offers a visual perspective on the timing and intensity of industrialisation across countries in the region.

Fig. 5.1

Industrialisation in Latin America. Data source: MOxLAD (MOxLAD, n.d.)

Taking Argentina as an example, the figure shows a gradual rise in the ratio beginning from the turn of the twentieth century, followed by an acceleration between the 1940s to the 1970s, with a decline and stagnation in the subsequent period. Brazil, Chile, Mexico, and Cuba show a roughly similar trend, rapid and deeper industrialisation in the mid-century period, followed by decline and stagnation. The decline after the 1970s peak has been attributed to premature de-industrialisation (Rodrik, 2016) but it also likely to reflect significant productivity improvements in agriculture, a measure of catch-up (Nin-Pratt & Valdés Conroy, 2020).

In Uruguay and Venezuela, and to some extent Panama, Peru, and Bolivia, the rise in industrial gross value added relative to agriculture is less pronounced, and peaks at a lower level. It is closer to an inverted ‘U’ than the inverted ‘V’ shown by the first group of countries. Central American countries, as well as Paraguay and Ecuador, show late and only very limited industrialisation. Costa Rica, on the other hand, shows a gradual but sustained rise in manufacturing value added relative to agriculture.

The Figure shows the ratio of industry to agriculture value added (in 1970 PPP USD). It captures the timing and depth of industrialisation across the countries in the region.

Comparison of industrialisation trends across the countries in the region point to three main country groupings: roughly a Central American cluster, an Andean cluster, and a Southern cone cluster plus Mexico and Costa Rica.⁴ In broad terms, long-period trends in industrialisation across these three groups of countries match trends in the timing and depth of the expansion of occupational insurance funds. By the end of the 1970s, Argentina, Brazil, Cuba, and Chile had well developed occupational insurance funds covering a substantial share of the labour force in dependent employment. Central American countries together with Paraguay and Ecuador, on the other hand lagged some way behind them. The influence of industrialisation on social protection appears dominant, but there are some cases that buck the trend. Uruguay and Costa Rica reached comparable high rates of worker participation in occupational insurance funds by the 1970s, while Mexico’s participation rates were significantly lower than predicted by its manufacturing value added relative to agriculture.

There is universal consensus in the literature in the central role of industrialisation and the associated development of state capacity in the expansion of social protection in the 1960s and 1970s (Mesa-Lago, 1978). Distinct industrialisation paths influenced the shape of emergent social protection institutions. In the literature on economic development in the region, a distinction is made between processes of industrialisation associated with implicit import-substitution in the earlier part of the twentieth century enforced by suspension of global trade, and the explicit state-led import-substitution policies embraced by some countries (Bértola & Ocampo, 2012). Industrialisation trends were evident in some countries in the region from the turn of the twentieth century (Williamson, 2011), a response to the suspension of world trade as a result of conflict. In some countries the state embraced a developmental model around import substitution policies (Astorga & Herraz-Loncán, 2020; Bértola & Ocampo, 2012). State-led import-substitutions policies reinforced and deepened the protection of favoured sectors in the economy. State-led developmental policies became central to the spread of occupational insurance funds for workers in these sectors. Social protection expansion through occupational insurance funds has important effects on employment. This is apparent from the fact that the initial focus on expanding social protection for urban workers and their dependants did not extend to other occupational groups such as informal or rural workers (Haggard & Kaufman, 2008). This became a feature of social protection in the region during the twentieth century. The sectoral affiliations of occupational insurance fund participants are fully consistent with state support for selected sectors in the economy. Occupational insurance funds lower competition in employment across and within the favoured sectors while raising competition for employment in the excluded sectors.

5.3 Limited Fordism and Social Protection

It is helpful to pay some attention to the qualitative as well as the quantitative quality of the employment effects of industrialisation.

Early industrialisation in the region involved a transformation of artisan manufacturing into larger-scale industrial production. This can be observed for selected countries from data collected for a 1966 Report on Industrialisation by ECLAC (ECLAC, 1966) which distinguishes between artisan and factory employment. Figure 5.2 shows the share of artisan employment and the share of factory employment in total employment in selected countries for the period 1925–1950, the early industrialisation expansion. In countries with a larger share of the labour force in industrial employment, like Argentina and Chile, expansion of factory employment reaches one half of manufacturing employment by 1950. In the other countries, starting from a lower share of employment in industry, factory employment is rising at a much lower pace and remained a minority in the later years. It is apparent from these data that employment in small enterprises was a feature of early industrialisation.⁵ In Latin America, industrialisation did not result in the agglomeration of employment in large Fordist enterprises to a significant extent.

Fig. 5.2

Artisan vs factory employment in selected countries 1925–1950. Data source: Statistical Annex to ECLAC Report on Industrialisation (ECLAC, 1966)

The Figure shows the artisan and factory shares of industrial employment for selected countries in Latin America. It shows the rise in factory employment against dominant artisan employment.

Fordist production has direct implications for social protection institutions. It relies on agglomerating workers with different skills within large integrated production processes. In combination with labour organisation and collective bargaining, fordist production can rapidly extend participation in social protection to a broad range of workers with different skill levels within enterprises (Rueda et al., 2015). In the social protection literature, Fordist production has a further, qualitative, dimension of interest, as the agglomeration of workers in large firms facilitates the development of labour organisations and enhances their political power. The limited expansion of Fordist production in Latin America might have prevented access to social protection institutions from broader categories of workers. The balkanised and selective occupational insurance institutions that emerged reinforced the stratified structure of employment. Fragmented institutions serving pockets of a highly stratified wage earner class is both a product of early patterns of industrialisation in the region and a causal factor in the fragmentation and stratification of employment since.

The limited spread of Fordist production in Latin America has been a constant in discussions on economic development and social protection. Eslava et al. (2021) provides a recent review of the literature and data on the significance of firm size for economic development in the region.⁶ They estimate that 18 percent of workers in the region are currently employed in firms with more than 100 employees, 14 percent in firms with 11–100 workers, and 36 percent in firms employing 10 workers or less. Own account workers account for a further 32 percent of employment.⁷ Firm size is also positively correlated with worker earnings and education levels and higher productivity when compared to smaller firms.

The small share of employment in large firms encourages a concern that the incidence of small firms and self-employment in Latin America, compared to other regions, is detrimental to productivity and might be a central explanatory factor in economic development outcomes (Levy & Cruces, 2021). This is presented as an issue of a ‘missing middle’: “…the size distribution of productive units in Latin America reveals a combination of massive excess fragmentation of production in the extreme left tail and an additional mass in large corporations, with a notable missing middle” (Eslava et al., 2021, p. 8). The connotation is that the distribution of productive units is also responsible in some way for structural inequality in the region and, by extension, its employment and social stratification. Hsieh and Olken (2014) dispute the ‘missing middle’ argument in development concerning firms size and employment. Whilst confirming the low absolute number of large firms, they demonstrate that the full distribution of firms in three middle income countries including Mexico is not unusual, it is well populated in the middle, and fails to show significant gaps.

The association with social protection participation is straightforward. Participation in occupational pensions is positively correlated with firm size, while participation in social assistance is negatively correlated with firm size. Social assistance programmes were explicitly designed to provide some protection to households dependent in informal employment while occupational pension plans were explicitly designed to include workers in larger firms and permanent employment relationship and exclude self-employed and rural workers.

The issue of causality is crucial to interpret these associations. It is a received view from the literature on this issue that changes in the structure of employment, and in legislation regulating the employment relationship, will be needed to secure a more comprehensive and inclusive social protection (Bertranou & Maurizio, 2011; Bosch et al., 2013). This is also a received view on the extensive literature in informality and the formalisation of informality. It is implicit in these views that causality runs from the structure of employment to social protection institutions.

On the other hand, Levy (2008) places the emphasis on the influence of social protection design on employment choices. In his view the combination of occupational insurance and social assistance undermines incentives for workers to opt for formal employment in larger firms. To the extent that workers can access health care and income protection from budget financed social assistance, the incentives to move to larger and more productive firms, forced to deduct pension fund contributions from their workers, are mitigated. Levy’s argumentation assumes a causal path going from social protection institutions to the distribution of employment across firms. Levy and Cruces (2021) develop a more sophisticated model in which the design of social protection institutions and firm growth are endogenous to multiple economy wide factors, via the incentives faced by economic agents. The causal links implied by Levy, from social protection design to firm size employment incentives, are now replaced by multiple cross-determining factors.

Their contribution raises a crucial policy dilemma, whether changes in the structure of employment are required to secure more inclusive and comprehensive social protection institutions (Martínez-Franzoni & Sánchez-Ancochea, 2017) or whether reforms to social protection institutions are required to reduce the stratification in employment (Levy, 2008). Proposals for delinking employment status and social protection participation suggest an alternative route to reducing employment stratification.

5.4 Earnings

Astorga (2017) constructs wage data for the ‘long century’ for three occupational groups (based on an ILO classification) for six countries. The occupational groups are: (i) relatively skilled workers (technicians and administrators); (ii) semi-skilled workers (semi-skilled blue collars workers and other urban workers in relatively low productivity sectors such as retailing and transport, and artisans); and (iii) unskilled workers (rural and urban unskilled workers, including domestic servants and street vendors). Figure 5.3 shows trends for monthly earnings in US$ at 1970 PPP for each of the countries and a simple regional average (LA6). Unfortunately, the countries included in Astorga’s dataset exclude Central American and Andean countries. Vertical dashed lines divide the ‘long century’ into three main periods 1900–1940, 1940–1980, and 1980–2011. These periods reflect early industrialisation, state-led development, and the neoliberal and post neo-liberal periods.

Fig. 5.3

Skilled, semi-skilled, and unskilled wage growth for selected countries 1900–2012. Data source: Astorga (2017)

The Figure shows long period monthly earnings of skilled, semiskilled, and unskilled workers for selected countries and a Latin America simple average, in 1970 PPP USD.

The 1900–1940 period shows only gradual improvements in earnings for the three occupational groups. Differentials across these groups are not marked. Frankema (2010) finds that inter-industry wage inequalities in Latin American countries were muted until the middle of the twentieth century. These findings are consistent with moderate differentials across workers grouped by skill. In the 1940–1980 period, the state-led developmental period in the selected countries, earnings rise sharply, although the rise is more moderate in Colombia, and to an extent Brazil. The post 1980 period is characterised by stagnation in labor earnings for all countries except Chile. There are cross-country differences in the intensity of the trend, but they share similar period patterns. Labour earnings show a marginal increase in the early industrialisation period, a sharp rise in the state-led development period, and stagnation in the neoliberal and post-neoliberal period.

Regarding differentials across the three skill groups, they show a relatively stable pattern in the early period, a sharper differentiation in the state-led industrialisation period, and some stability in skill differentials in the neoliberal period, apart from Chile who shows widening skill earnings differentials. The timing of industrialisation could explain variations in the pattern across countries. Earlier industrialisation in Argentina would probably explain the greater intensity of the earnings growth for the skilled group. The state-led development period shows a rapid increase in skilled and semi-skilled earnings reflects rapid industrialisation in protected sectors. Venezuela in particular shows a very significant rise in earnings for the skilled group. The last period is one of wage stagnation for all workers in all countries except Chile and to a lesser extent Brazil.

For the full period, and focusing on the LA6 data, wage earnings growth can be observed for all groups of workers with a widening in differentials. According to Astorga, the wages of the unskilled group rose in the full period by 147 percent while the average wage growth was 257 percent. The growth in earnings appears to have petered out in the neoliberal period. The average earnings for the LA6 countries at the end of the data period in 2011 are lower in PPP terms than at the end of the state-led period in 1980.

The expansion of occupational insurance funds in the industrialisation period will have contributed to the rise in earnings differentials, through a relatively stronger rise in earnings of skilled workers. Occupational insurance funds restricted competition within the sectors privileged by import-substitution-industrialisation. The exclusion of rural and unskilled workers expanded competition among these workers. Occupational insurance funds reinforced sectoral stratification. The fragmentation and selectivity associated with occupational insurance funds in Latin America enforced significant intra-wage earner divisions, as shown by growing skill differentials.

Efforts to undermine workers organisations and minimise their influence over social protection policies and institutions in the neoliberal period together the acute economic and financial crises that accompanied structural adjustment helped moderate earnings differentials but reinforced employment stratification. The expansion of social assistance and the rise in real minimum wages in the new century led to a relative improvement in the earnings of low income workers and raised reservation wages. These policies led to a reduction in wage earner stratification. Duality in social protection provision contributed to sustain wage earners stratification by skills.

The sustained reduction in labour earnings inequality in Latin America since the turn of the century has attracted considerable attention. Messina and Silva (2018) find that labour earnings inequality decrease sharply in the first decade of the 2000s in 16 out of 17 countries they study. Faster rise in the earnings of lower paid workers is the main driver for the overall reduction in wage inequality. The extensive literature seeking to identify the causal factors responsible for the relative improvement in the earnings of less skilled workers has explored a wider range of possibilities. They range from supply side factors associated with education expansion; demand side changes in global demand and improvement in the terms of trade; and policy factors such as improvements in the minimum wage (Firpo & Portella, 2019). Supply side explanations emphasise the expansion of secondary and tertiary education in the last quarter of the twentieth century which might have reduced the skill premia observed in the 1980s and 1990s. The commodity boom brought about by the integration of China in the world economy and its rapid growth could have raised demand for workers in primary and tertiary sectors, with positive implications for their earnings. A widespread rise in real minimum wages and the expansion of conditional income transfers could have lifted reservation wages among less skilled workers. The literature has not reached agreement. Interestingly, Messina and Silva conclude that “most of the decline is explained by falling intra-group wage inequality” (Messina & Silva, 2018, p. 61), effectively discounting the influence of changes in worker characteristics or sectoral employment.

5.5 Tertiarisation and Social Protection: Driver or Consequence?

In the second half of the twentieth century tertiarisation trends began to dominate the structure of employment. Secular reductions in the share of employment in agriculture and stagnation in industrial employment resulted in a rising share of employment in services as shown in Fig. 5.4. What are the implications for the link between social protection and employment?

Fig. 5.4

Share of employment by industry sector. Data source: Economic Transformation Database (de Vries et al., 2021)

The Figure shows trends in employment shares by main industrial sectors: agriculture, industry, and services from 1950 to 2014. It captures the secular decline in the share of employment in agriculture, and the rise in the share of employment in services. The share of employment in industry has retained stability through the period.

Taking for granted that the direction of causality goes from employment to social protection, researchers have argued that the expansion of tertiary employment was in part responsible for the stagnation of social protection in the neoliberal period. Compared to industrial employment, tertiary jobs outside public services are less likely to agglomerate large numbers of skilled workers. They are less likely to rely on firm specific skills and therefore long-term employment relationships. Most firms in the tertiary sector rely largely on general skills and are not dependent continuous employment. They likely see the costs of occupational pension funds as non-recoverable expense. Researchers have noted the more volatile nature of services employment is ill suited to occupational insurance funds, which sustainability and resilience relies on more permanent working relationship. Tertiarisation is, from this perspective, an explanation for the stagnation of participation in occupational insurance funds.

Approaching this from the perspective that social protection influences employment provides a different, and perhaps more robust, reading of tertiarisation. Social protection reforms are capable of driving change in the nature of the employment relationship itself. The introduction of individual retirement savings plans supports a more flexible approach to employment relationships. Shifting social protection to individual retirement savings plans would encourage firms and workers to restructure their employment relationship.

Occupational insurance funds require active management on the part of employers with associated costs. The economic literature has examined the recruitment and retention effects of occupational insurance funds (Lazear, 1983). Pension provision designed for the armed forces is a good example. Occupational pension plans in the armed forces require many years of continuous contribution before allowing access to entitlements. This is intended to incentivise recruits to work through the costs of training. Occupational pensions with long vesting periods select workers with long-term horizons, prepared to trade off current pay for future pension income. Firm-specific skills training is unlikely to be attractive to workers who move jobs, while general skills can provide such workers with substantive benefits.

Shifting social protection to individual retirement savings plans enables firms and workers to restructure their employment relationship (Barrientos, 1998b, 1998a; Barrientos & Firinguetti, 1995). Individual retirement savings plans could more easily accommodate a measure of volatility in employment, due to their portability across jobs and the individualised entitlement rules. Participants get what they save, plus the returns on their savings. Volatility of employment is unlikely to undermine the operation of savings plans, especially where fees are extracted on a regular basis form the balances of the plans.

Occupational insurance funds can help firms deploy productivity incentives for workers in activities with costly monitoring. The fragmented nature of tertiary employment and the large share of small firms in the private sector implies that the monitoring of worker productivity is costless for employers. A shift to individual retirement savings plans enables changes in the employment relationship in firms benefiting from costless monitoring of worker effort. In sum, pension reforms facilitate tertiarisation.

The stagnation in industrial employment also bears some consideration. Discussing trends in industrialisation in low and middle-income countries, Rodrik (2016) argues that premature de-industrialisation is likely to have a stunting effect on the economies and institutions of these countries. Late and weak industrialisation in Latin American countries, compared to early industrialisers, might help explain the relative weaknesses and fragmentation of early welfare institutions. Taking the argument forward in time, Rodrik argues that de-industrialisation, associated with the shift to export-led growth in the last quarter of the last century, has further undermined the emergence of strong and unified welfare states. Rodrik is perhaps over-pessimistic about the extent to which weak and late industrialisation might permanently affect late industrialisers. He writes: “Historically, industrialisation has played a foundational role in creating modern state and democratic politics. Its relative absence in today’s developing societies could well be the source of political instability, fragile states, and illiberal politics” p. 24. His argument relies explicitly on applying the trajectory of early industrialisers to countries in Latin America and elsewhere. He takes it for granted that the direction of causality goes from employment to social protection.

However, data on selected countries in the region shown in Fig. 5.4 do not fully support the de-industrialisation argument. The data confirm that the share of industrial employment has been relatively stable through the period. At any rate, relatively higher productivity in industry than in services or agriculture is consistent with a moderate fall in industrial employment over time, as productivity gains materialise.

5.6 The Knowledge Economy and Skill Polarisation

There is growing interest on the likely impact of the ‘knowledge economy’ on employment in the region. Undoubtedly, the spread of information technology and of innovations in automation and communications have had a significant impact on the way we work and on workers skills set requirements. In early industrialisers, research centres around occupational polarisation caused by off shoring and automation. Off shoring describes the relocation of economic activity to emerging economies and consequent effects on job demand. Automation describes the spread of computer assisted machine production as a complement or replacement of workers. The combination of off shoring and automation in early industrialisers has affected particularly routine occupations because they are more susceptible to automation. This has resulted in a hollowing out of the employment structure, particularly in manufacturing and clerical jobs. The implications for labour demand are a polarisation of employment, sustained demand low skill and high skill workers contrasting with a collapse in demand for routine blue and white collar employment (Autor & Dorn, 2013).⁸

This issue is relevant for countries in Latin America. First, research into off shoring and automation in early industrialisers has generated new insights and perspectives on the relationship between skills, capital, and employment. It has led to the development of a novel task approach to employment (Acemoglu & Autor, 2011; Autor, 2013). Second, off-shoring and automation raise an interesting set of questions around the potential consequences for employment in the region. Are Latin American countries likely to benefit from the jobs off shored to the periphery. Will this raise demand for routine employment? Will demand for highly skilled, and low skilled jobs drain our economies of their employment? Are there any new investment opportunities that could impact on employment structures in the region? And what is the contribution of social protection to this change?

Research on the empirical effects of off shoring and automation on the employment structure in Latin America has not reached clear-cut conclusions. One strand of research has focused on assessing the exposure of occupations to potential routinisation. Gasparini et al. (2021) study the routine content of jobs in six countries in the region to establish how open they are to job polarisation. They find that unskilled occupations have a higher routine content compared to semi-skilled occupations and are significantly higher than for skilled occupations. They conclude: “[A]lthough there is considerable heterogeneity across countries, the general result holds: in all six countries employment has increased less, or even decreased, in occupations with high routine task content. As a result of the asymmetric changes in employment, the overall degree of routinization has mildly decreased in the Latin American economies over the last two decades. The fall was more marked in the 2000s than in the 2010s. Between and within-sector changes are relevant to understand these patterns.” (Gasparini et al., 2021, pp. 2–3). This finding is confirmed by research in Bonavida et al. (2021) for a range of Latin American countries and Nogueira (2015) for Brazil.

Instead of working out the potential scope for change, we could attempt to measure actual changes in employment by occupation using ISCO08 data from the ILO covering 11 Latin American countries.⁹ This classification distinguishes occupations according to their skill content (Ospino Hernandez, 2018). The nine one-digit occupational ISCO groups can be further consolidated into three: routine (clerical support workers, skilled agricultural, forestry and fishing workers, craft and related workers, plant and machine operators and assemblers); non-routine manual (services and sales workers and armed forces); and non-routine cognitive (Managers, Professionals and Technician and associate professionals). Figure 5.5 shows the ratio of the share of workers in each of the three categories in the last year that data is available for the relevant countries compared to the initial year.

Fig. 5.5

Job Polarisation? Share ratios of cognitive, routine, and non-routine manual occupations for selected countries. Data source: Own analysis of ILO ISCO08 data file accessed 21/9/21

The Figure shows the ratio of the shares of employment in non-routine cognitive, non-routine manual and routine occupations for selected countries. Ratios greater (less) than 1 show growth (decline) in occupations. It captures potential employment polarisation between the first and last available in the data.

Job polarisation of employment due to automation would be consistent with growth in the share of employment in non-routine cognitive jobs and possibly some growth in the share of non-routine manual jobs (to the extent that high skill workers need complementary ancillary workers, e.g., home deliveries and take away meals). In contrast, the effect of a transfer of routine jobs from early industrialisers due to off shoring would be consistent with a rise in the share of routine jobs.

Central American countries and Ecuador show an increase in the share of routine jobs as well as declines in the job shares of both non-routine manual and non-routine cognitive jobs. The same trend can be observed in Bolivia, but the changes are marginal. Argentina is an intriguing case because of the large increase in the share of routine job share. This places Argentina closer to the trend for Central American economies. Mexico shows a rise in both routine and non-routine cognitive job shares, together with a decline in the share of non-routine manual jobs.

Uruguay, Chile, and Brazil exhibit some signs of job polarisation due to the growth in the share of non-routine cognitive job share. However, this is combined with some decline in the share of routine and non-routine manual job share. Perhaps more appropriately, these countries demonstrate the impact of the’ knowledge economy’, un upgrade in the share of workers performing skilled jobs.¹⁰

It is hard to see a common trend across the Latin American economies included. Southern cone countries, less Argentina plus Mexico, show signs of the ‘knowledge economy’ restructuring employment, in the rise of non-routine cognitive job share. For Central American and Andean countries, occupational changes are in line with off shoring of routine jobs from early industrialisers.

The off shoring and automation narratives emphasise demand side factors in explaining occupational change in early industrialisers. In the context of Latin America, it is interesting to consider supply conditions. In the last quarter of the twentieth century Latin American countries upgraded their education provision (Lindert, 2010, 2021).

How well positioned are Latin American countries to meet the coming knowledge economy? Figure 5.6 cross tabulates the share of the relevant population group enrolled in tertiary education with the share of non-routine cognitive jobs across several Latin American countries. Paying attention to supply conditions does not change the findings from the examination of occupational employment changes in significant ways. Countries that experienced a rise in non-routine cognitive jobs also show higher than average share of tertiary enrolments in the relevant age group. Aside from Argentina, countries showing increases in the share of routine jobs also have lower than average share of tertiary enrolments in the relevant population. Among these countries, supply side constraints match weak demand trends.

Fig. 5.6

The coming ‘knowledge economy’? Data source: Tertiary enrolment from Our World in Data; Non-routine cognitive share of employment from own estimation using ILO ISCO08 data file accessed 21/9/21. Data points are for 2014, except non-routine cognitive share for Chile and Dominican Rep. in 2015 and for Peru in 2010

The Figure shows a crosstabulation of the share of tertiary enrolments and the share of non-cognitive jobs in employment for selected countries. It offers a graphical summary description of supply and demand conditions in preparation for the ‘knowledge economy’.

Individual retirement savings plans and social assistance have features that will contribute to the changes in employment associated with automation and with job relocation. As discussed in Chap. 3, both individual retirement savings plans and social assistance expansion are associated with a shift in dominant stratification from sectors to skill. Changes in employment due to automation and job relocation are part of this shift.

Occupational insurance funds provided incentives for employment in the favoured sectors of the economy. They reduced competition within these sectors and restricted job mobility across sectors. High replacement rates for stayers in occupational pensions were dependent on limited benefits for lower paid workers with low probability of meeting vesting periods and on public subsidies. The introduction of individual retirement savings plans and recent reforms to occupational pensions (Altamirano Montoya et al., 2018) will have the effect of raising competition and job mobility within and across sectors. Deterioration or scaling down of subsidised replacement rates in occupational insurance funds, due to financial pressures and reduced levels of public subsidies or internal subsidies, will further reduce incentives for employment in routine non-manual jobs, their core constituency.

Social protection reforms in Latin America since the beginning of the neoliberal period have supported a shift to skill stratification. As occupational insurance funds stagnated or declined, individual retirement savings plans were pushed into the forefront. Individual retirement savings plans are not sector specific or sector selective. Individual retirement savings plans lack redistributive features with entitlements based on capacity to save rather than group membership. Individual retirement savings plans replace sector stratification with skill stratification.

From the turn of the century, social assistance has expanded its reach of low-income groups often dependent on informal or unprotected employment. The fact that social assistance entitlements are not dependent in employment status removes any sector stratification bias. At the same time, their social investment orientation, if effective, raises education and skill levels among younger cohorts and will upgrade the skill composition of the labour force (Parker & Vogl, 2018).¹¹ Social assistance transfers, and recent minimum wage upgrades, have the effect of raising the reservation wage of lower skilled workers, especially in routine manual occupations. The expansion of social assistance can offer support for changes in the employment structure associated with automation and job relocation.

Given conditions in the region, a ‘knowledge economy’ scenario is consistent with current social protection trends: the stagnation and decline of occupational insurance funds; a realignment of social protection and skill stratification through individual retirement savings plans, and the expansion and institutionalisation of social assistance supporting non-routine manual workers.

5.7 New Forms of Integration with the World Economy

Global production networks have far-reaching implications for employment in Latin America, especially in the context of the commodity boom (Ocampo, 2017; Pérez, 2010). Global production networks, also referred to as global value chains, describe agreements involving firms in different countries to contribute specific components of a production line. They involve the subdivision of production processes and the allocation tasks across national borders. A key characteristic of global production networks is that component firms are not vertically integrated with the dominant firms, their collaboration is based on mutual advantage not ownership. This process is the reverse of multinational corporations’ vertical ownership and command. Component firms are in a network but retain a measure of autonomy and are fully inserted in their domestic economy.

This is related to off shoring in early industrialisers where firms relocate the production of components to locations where conditions are more favourable (say due to labour supply or natural resources), or at least less unfavourable (say locations with less costly environmental and labour regulations).

The first issue to explore is the extent of Latin American countries’ participation in global production networks. A measure of global production network participation can be constructed using UNCTAD-Eora data (Casella et al., 2019). It estimates flows of value added in exports, extracted from country level input-output matrices, aggregated for the global economy. A measure of global production network participation subtracts backward value added from forward export value added. Forward export value added is the value added in goods and services used as intermediate inputs in third countries and excluding value added in goods exported for direct consumption. Backward value added is the value added in imports associated with global production network in the home economy (Casella et al., 2019). Figure 5.7 shows the trend in global production network participation as fraction of exports for Latin America countries in the database 1990–2018.

Fig. 5.7

Global Production Networks participation index. Data source: UNCTAD-Eora data (Casella et al., 2019). Index adds forward and backward linkages

The Figure shows the Global Production Networks participation index for Latin American countries. The index is the share of export value added accounted for by forward and backward linkages.

The general trend is concave over time, with participation rising in the 2000s but showing a gentle stagnation or decline after 2010. Global value network participation grew in significantly in the 2000s. The 2008 global financial crisis, with a delayed impact in Latin American countries, appears to have arrested the rise in global production network participation.¹² Among the largest economies of the region, global production network participation remains at significant levels. The fraction of export value added accounted for by participation in global production networks peaked at around 0.5 for the countries with the highest levels of participation, namely Chile and Peru. For Argentina, Bolivia, Brazil, Costa Rica, and Uruguay global production network participation peaked around 0.4.

The next issue is to assess the contribution of participation in global production networks in employment. LACEX jobs data supports measures of the direct number of jobs engaged in exports (forward linkages) as well as the indirect number of jobs (backward linkages) associated with domestic inputs into the sector exports (Cali et al., 2016).

Figure 5.8 shows the share of jobs estimated to be associated with global production networks as a share of employment. Where data are available, the Figure compares early and late observations for each country. Care should be exercised in interpreting the figures for 2011 as they probably reflect the delayed effect of the 2008 global financial crisis on the region. A rough average for the region suggests that around one sixth of all jobs are associated the global production networks integration into the global economy. Costa Rica is an outlier with close to one half of all jobs contributing to exports. This exercise demonstrates that the integration of Latin American economies into the global economy, and perhaps the significance of participation in global production networks goes much further than measures of direct jobs associated with exports alone. They substantially understate a country’s job dependence on exports.

Fig. 5.8

Share of direct and indirect jobs embedded in export sector. Data source: LACEX database (Cali et al., 2016)

The Figure shows the share of jobs embedded in the export sector, including those directly and indirectly engaged, for selected countries and years.

Social protection reforms in the region since the neoliberal period have facilitated global production networks by supporting skill stratification. The discussion in the previous section regarding the contribution of social protection to changes in employment applies here in full. Individual retirement savings plans facilitate participation in global production networks due to their flexibility with regards to job mobility and reliance on individual savings capacity. Social assistance provision also facilitates employment in global production networks by lower skilled workers. The limited role and responsibility of employers in the management and financing of individual retirement accounts and in the budget-financing of social assistance minimise the costs associated with compliance with social protection labour regulations and enhances their competitive position. Social protection reforms support the temporal and changing nature of firms’ global production networks commitments.

5.8 Conclusions

This chapter examined the association existing between social protection institutions and employment. Researchers have paid a great deal of attention to the close connection between social protection institutions and employment. The definition of informal employment in common use, namely workers who lack social protection, underlines the close association between them.

Consensus over the significance of this association breaks down when the issue becomes to establish the direction of causality. The conventional approach, supported by current theories of welfare institutional development, is to assume that the structure of employment shapes social protection institutions. This makes sense when researching the origins of welfare institutions. Social protection institutions in Latin America emerged with changes in employment imposed by industrialisation. The emergence of an urban working class for whom employment was binary posed large risks to the wellbeing of workers and their dependents in the event of unemployment, sickness, or old age. Government supported occupational insurance funds emerged with the incorporation of urban industrial workers in mass politics. However, once social protection institutions are in place, they exercise considerable influence on employment. The direction of causality is now from social protection to employment. The discussion in the chapter explored the stratification effects of social protection institutions on the structure of employment. Adopting this approach new insights into social protection.

Occupational insurance funds expanded in the region with state-led import-substitution-policies. They reinforced the dominance of sectoral stratification by enhancing worker incentives for employment in the protected sectors are reducing competition within selected sectors. Occupational insurance funds strengthened recruitment and retention within these sectors. They also excluded workers from outside these sectors. Core features of occupational insurance funds, like vesting periods, redistributed entitlements away from low skilled workers with discontinuous employment and towards skilled workers with continuous, long tenured, employment.

Pension reform in the 1990s, including individual retirement savings plans, and the growth of social assistance represent a shift towards skills stratification. Individual retirement savings plans are not directed to sustain favoured economic sectors, but instead aim at enhancing competition within and across sectors, job mobility, and to minimise redistributive features that weaken the association between entitlements, skills, and savings capacity.

The expansion of social assistance in the new century further reinforces a shift towards skill stratification. In this case, they are directed at low paid workers in informal employment. Social assistance entitlements are independent of employment status and sectoral affiliation. Conditional income transfers’ social investment orientation is aimed at upgrading educational attachment and at promoting the acquisition of skills among participant children. Conditional income transfers seek to improve the productive capacity of low income groups in the medium and longer term.

The chapter discussed social protection contribution to three changes in employment in the region: tertiarisation, job polarisation and relocation, and participation in global production networks. It argued that social protection reforms, individual retirement savings plans and social assistance, have contributed to these changes in the employment structure. Approaching the association between social protection and employment as a causal path brings new insights on social protection institutions in the region.