Open-access Drivers of employment change in Brazil in sectors by technological intensity: a structural decomposition analysis

Drivers de mudança do emprego no Brasil em setores por intensidade tecnológica: uma análise de decomposição estrutural

Abstract

This paper investigates the drivers of formal employment growth in Brazil during the 2010s, exploring the heterogeneity of human capital and technological intensity across sectors. Using input-output matrices, the study performs a structural decomposition of employment, identifying the contributions from technological changes, labor intensity, and final demand structure in different subperiods. The results indicate that the new economic factors affected less-skilled workers more. Increased labor productivity and economic changes led to higher unemployment, mainly for these workers. On the other hand, the growth in final demand driven by household consumption was the main factor behind creating new jobs, especially for more skilled workers, suggesting changes in the formal labor market with a qualification bias. Moreover, government consumption, exports and investments also contributed to creating new job opportunities.

Keywords: employment; human capital; technology; input-output; Brazil

Resumo

Este artigo investiga os drivers do crescimento do emprego formal no Brasil durante a década de 2010, explorando a heterogeneidade do capital humano e a intensidade tecnológica entre os setores. Utilizando matrizes de insumo-produto, o estudo realiza uma decomposição estrutural do emprego, identificando as contribuições das mudanças tecnológicas, da intensidade do trabalho e da estrutura da demanda final em diferentes subperíodos. Os resultados indicam que os novos fatores econômicos afetaram mais os trabalhadores menos qualificados. O aumento da produtividade do trabalho e as mudanças econômicas levaram a um maior desemprego, principalmente para esses trabalhadores. Por outro lado, o crescimento da demanda final impulsionado pelo consumo das famílias foi o principal fator por trás da criação de novos empregos, especialmente para os trabalhadores mais qualificados, sugerindo mudanças no mercado de trabalho formal com um viés de qualificação. Ademais, o consumo do governo, as exportações e os investimentos também contribuíram para a criação de novas oportunidades de emprego.

Palavras-chave: emprego; capital humano; tecnologia; insumo-produto; Brasil

1 Introduction

Driven by improved macroeconomic conditions and the expansion of exports, the Brazilian economy entered the new century reversing the downward employment trend observed after the economic opening in the 1990s (Cunha et al., 2021; Macedo and Porto, 2021). Between 2003 and 2008, export demand was the main driver of the country's growth, with an average annual growth rate of 7.7%. After the global crisis of 2007, the Brazilian government implemented countercyclical policies to maintain economic evolution. This intervention resulted in a broad process of formalizing employment contracts, increasing the population's purchasing power, and reducing socioeconomic inequalities. In parallel, public investments and productivity improvements also contributed to the Brazilian economy advancing at an average rate of 3.5% per year until 2014 (Magacho and Rocha, 2022; Macedo and Porto, 2021). As a result, between 2003 and 2014, formal employment jumped from 29.5 million to 49.5 million, a remarkable increase of 67.8%. The number of workers who had completed high school and higher education also grew considerably (over 80%), reaching a total of 22.8 million and 9.3 million in formal employment, respectively.

However, starting in 2014, the Brazilian economy began to lose steam and experienced a significant slowdown in economic growth. Although there is no consensus on the exact causes of this economic crisis, it is widely accepted that it was the result of a combination of bad government interventions, lasting shocks to demand and supply, and a loss of government credibility in conducting monetary and fiscal policies (Bastos, 2017; Barbosa, 2017).1 Between 2014 and 2016, the country's average growth rate fell to an alarming -2.3%. Although export demand has evolved in the opposite direction, with a 2.3% increase, the agro-export model presented less potential to sustain production and employment growth (Macedo and Porto, 2021; Magacho and Rocha, 2022). In 2017, the unemployment rate reached levels of a depressed economy (13.2%), equivalent to 13.5 million unemployed people. Not surprisingly, informal work, which had previously decreased, grew to 11.2 million in 2018 (Nassif, 2017).

This paper aims to identify the drivers of formal employment growth in Brazil during the 2010s. The periods of economic growth (2010-2014) and recession (2014-2018) are considered to carry out a comparative analysis of the effects of different factors on the variation of employment by the level of human capital and technology intensity of the productive sectors. The technological level is defined by a sectoral classification developed by the OECD based on the intensity of investment by industries in research and development (R&D). This classification is used to group the sectors of input-output matrices according to their level of technological intensity (Morceiro, 2018; Galindo-Rueda and Verger, 2016). Using the input-output model with sectors by technological intensity as a framework for structural decomposition analysis, this study seeks to identify the determinants of employment growth by the level of human capital, considering general equilibrium effects. This method allows for estimating the direct and indirect impact of variations in each demand component and their potential impacts on employment variations due to multiplier effects. It also enables the estimation of the effect of increased labor productivity and technological changes on employment.

By integrating endogenous factors of economic growth into input-output modeling, it is possible to analyze the Brazilian labor market from the perspective of the heterogeneity of human capital and technological intensity of sectors. This allows for a more detailed examination of the impacts of economic changes on the labor market (Acemoglu and Autor, 2011; Autor et al., 2003; Romer, 1990). Sectors with higher technological intensity usually require more skilled labor and have higher productivity and wages. On the other hand, sectors with lower technological intensity tend to have lower productivity and wages, but job availability is more elastic for less skilled workers (Compagnucci et al., 2021; Acemoglu, 2002). Examining employment variations across sectors by technological level is also essential for identifying economic growth opportunities and reducing unemployment. Sectors with higher technological intensity often have greater productivity growth and innovation potential, which can lead to job creation and economic development (Hauknes and Knell, 2009).

Another possibility is to identify ways to promote economic growth and, at the same time, ensure social inclusion. Sectors with lower technological intensity tend to offer more opportunities for less skilled workers, providing them with a source of employment and income. Considering the differences between sectors, it is possible to inform economic planners which sectors offer more potential for this class of workers, enabling them to develop programs to facilitate their entry into the formal labor market (Lima et al., 2020; Glomm and Ravikumar, 1992). Finally, sectors with higher technological levels tend to demand more qualified labor, requiring investments in training and education programs to ensure the availability of workers with the necessary qualifications to fill job positions (Acemoglu, 2022; Acemoglu and Restrepo, 2019).

Previous studies have also used input-output modeling to analyze labor market patterns and identify the factors driving employment changes in different economic contexts. Gregory et al. (2001) observed that demand growth and technological advances led to a significant skill bias in the British labor market. Tin (2014) revealed that the main drivers of employment growth in the manufacturing sector during Malaysia's industrialization were domestic and external demands. Carrascal Incera (2017) found that during periods of economic growth in the European Union, household consumption and exports were the main sources of job creation for young people, while productivity gains were the main source of unemployment. Madariaga (2018) verified that during Spain's economic boom, demand changes were a key factor in employment growth in sectors and occupations requiring more professional skills, while labor productivity gains partially offset this effect. Doan and Long (2019) also found that domestic and export demand expansion drove employment growth in China. Barba and Iraizoz (2020) further identified that internal demand was the engine of employment growth for women in the European Union, especially in public services sectors, while labor intensity had the opposite effect, mainly in private services sectors. These changes led to a decrease in gender segregation levels in productive sectors.

In Brazil, input-output modeling and structural decomposition analysis are increasingly used to examine labor market changes. Sesso Filho et al. (2010) observed that trade liberalization led to a shift in employment from the agricultural and industrial sectors to the trade and service sectors. Perobelli et al. (2016) found that final demand was the main driver of formal job creation across almost all levels of education during the post-commercial opening period, while productivity had an inverse effect. Fiuza-Moura et al. (2016) verified that technological changes during the 2000s created service sector jobs and eliminated commerce jobs. Catelan et al. (2021) found that the decrease in youth employment in the Brazilian economy resulted from an increase in overall labor productivity and a reduction in this factor's utilization level, compensated by increased household consumption, investment, exports, and government spending. Acypreste (2022) observed that, although technology increased unemployment in the Brazilian economy between 2000 and 2015, these losses were more than compensated by creating new jobs due to demand growth.

This paper contributes to the presented literature in several important ways. Firstly, it identifies the drivers of variation in formal employment in the Brazilian economy, considering endogenous factors that promote economic development, such as workers' levels of human capital and the technological intensity of productive sectors. To our knowledge, this empirical investigation is unprecedented for the Brazilian labor market. It builds on consolidated evidence showing that workers and sectors are heterogeneous regarding skills, knowledge, and technology. Consequently, workers and sectors are differently affected by economic transformations, such as those occurring during periods of economic growth and recession, as well as those resulting from technological innovations such as production automation, artificial intelligence, and robotics (Lise and Postel-Vinay, 2020; Acemoglu and Autor, 2011; Hauknes and Knell, 2009). Secondly, considering the analyses already done for the Brazilian economy, this empirical study advances the literature by disaggregating final demand to identify the role of household and government consumption, gross fixed capital formation (investment), and exports in the variation of formal employment. Finally, this analysis provides a comprehensive understanding of changes in the labor market during the 2010s. In particular, it contributes to assisting in the formulation of policies aimed at resuming formal employment growth in Brazil.

The remainder of the paper is structured as follows. Section 2 presents the methodology. Section 3 presents the empirical results for formal employment by level of human capital and technological intensity of the sectors for Brazil. Finally, Section 4 provides a general discussion and the main conclusions.

2 Methodology

2.1 The structural decomposition model

Structural decomposition analysis (SDA) is based on intersectoral internal and external flows of inputs and outputs and demand model specifications to capture direct and indirect demand effects in all sectors. The basic structure of the SDA model consists of an input‒output matrix, which represents the economic system with (n) productive sectors by technological intensity (Rose and Casler, 1996; Dietzenbacher and Los, 1998). This matrix denotes payments in monetary units between sectors in year (t), as well as intermediate inputs (xij) allocated in the production process of other goods. Columns (j) denote sectors that demand inputs, while rows (i) represent sectors that supply goods and services. Final demand (D) encompasses household consumption (C), government expenditure (G), gross fixed capital formation (investment) (K), and exports (E).2 Thus, the gross value added of each sector is determined by the sum of intermediate consumption and final demand (Barba and Iraizoz, 2020). From this structure, the Leontief demand model is expressed as follows:

X=(IA)1D (1)

where X is a vector (6×1)of sectoral production by technological intensity, I is an identity matrix (6×6), and A is a matrix (6×6) of technical coefficients, where each element aij is expressed by a fixed proportion of inputs used in the production process (xij/Xj), while (IA)1 is the inverse Leontief matrix (L) with direct and indirect technical coefficients. These coefficients indicate, in monetary units, the production of sector (i) necessary to meet one unit of final demand. The above model can be simplified as X=LD.

To analyze changes in the labor market, five vectors of employment coefficients are constructed for each year tϵ{2010,2011,,2018} by level of human capital, according to the following equation:

hiq=HiqXi (2)

where the vector of employment by level of human capital (hiq) is a proportion of the number of formal jobs at each level of qualification (Hiq) by the gross value of production of each sector by technological intensity (Xi). The subscript qϵ{1,2,,6} corresponds to formal jobs by level of human capital in sectors by technological intensity (i). The levels of human capital are categorized as follows: low (1), low-medium (2), medium (3), medium-high (4), high (5), and total (6). The sectors by technological intensity are expressed by the subscript iϵ{1,2,,6}, with the first five being the sectors by technological intensity categorized similarly to human capital, while the sixth represents other sectors.3

Next, a matrix of employment generation by level of human capital is constructed, according to Eq. (3). To measure it, first, the matrix ˆhq is created, which is a diagonal matrix calculated based on the vector hq. By multiplying the new matrix by the Leontief inverse, the matrix G(ˆh) is found (Perobelli et al., 2016):

G(ˆhq)=ˆhqL (3)

The sum of the elements in each column of the matrix G(ˆh) equals the simple employment multiplier (employment generator) by human capital level (q) for each sector by technological intensity (i).4

Given the objective of this empirical study to analyze the drivers of changes in employment based on the level of human capital, we consider the employment variation matrix (ΔH) given by Eq. (4).

ΔH=HtH0=ˆhtLtDtˆh0L0D0 (4)

Where 0 and t represent the initial and final years of each subperiod, including the initial and final periods of the study. For this purpose, the periods from 2010 to 2014, 2014 to 2018, and 2010 to 2018 are considered.

The structural decomposition of employment variation for each subperiod follows the modeling of Barba and Araizoz (2020), Madariaga (2018), and Carrascal Incera (2017). In this context, for structural decomposition, employment variation vectors were constructed by the level of human capital as follows:

Δˆh=ˆhqtˆhq0,withq=1,2,...,6. (5)

where qcorresponds to formal employment by level of human capital, where q corresponds to formal employment by level of human capital. Combining Eqs. (4) and (5), we can initially perform the structural decomposition of employment into three components using the polar average approach of Dietzenbacher and Los (1998)5:

ΔH=12(Δˆh)(LtDt+L0D0)Labor+12(ˆhtΔLD0+ˆh0ΔLDt)Technology+12(ˆhtLt+ˆh0L0)(ΔD)FinalDemand (6)

Eq. (6) shows that the variation in employment between two periods is explained by changes in labor intensity, technology, and final demand structure (Barba and Araizoz, 2020). In this framework, technological changes represent any factors that cause changes in technical coefficients, such as technological innovations, technical substitution, and economies of scale (Rose and Casler, 1996).

For a better understanding of the determinants of formal employment variation, the final demand (D) can still be broken down into different components, such as household and government consumption, gross fixed capital formation (investment), and exports - Eq. (7). From here on, for the sake of simplicity, we will refer to gross fixed capital formation as “capital” or “fixed capital” (investment).

Finaldemand=12(ˆhtLt+ˆh0L0)(ΔDFHousehold+ΔDGGovernment+ΔDKCapital+ΔDEExport) (7)

This same mathematical framework is used to perform the structural decomposition of employment at the aggregate and sectoral levels, aiming to present a more comprehensive overview of changes in the formal labor market in Brazil.

2.2 Structural change index

The Structural Change Index (SCI) estimates the reallocation effect caused by various factors that influence employment and production in the Brazilian economy, such as technological transformations, foreign trade, and changes in domestic demand. The SCI is calculated as follows:

SCI=12|pitpit1|,with0SCI100 (8)

where pit and pit1 represent the share of each sector in the total number of jobs in the economy in different periods (t). The use of the absolute value ensures that positive and negative values are not nullified when summed, while the sum is divided by two to avoid double counting. The SCI can range from 0% (no structural change) to 100% (complete structural change). Overall, the closer to 100%, the greater the structural change in the sectors’ participation in the formal employment of the economy. That same calculation is made for gross sectoral production (Sesso Filho et al., 2010).

2.3 Data

Input-output matrices (IOMs) were used to perform the structural decomposition of formal employment growth in Brazil from 2010 to 2018. The period is limited due to the unavailability of comparable input-output matrices for the Brazilian economy for years after 2018. These matrices are generated from information provided by the Sistema Nacional de Contas (SNC) of the Instituto Brasileiro de Geografia e Estatística (IBGE), following the method of Guilhoto and Sesso Filho (2010).6 The method involves a procedure for combining information from the tables of resources and uses of goods and services of the national accounts. These matrices contain statistical information on production and intermediate consumption in monetary units for 128 products and 68 economic sectors. Next, we aggregate these economic sectors by technological intensity following the procedure below and adjust the monetary values to the 2010 base year using the implicit GDP deflator (single deflation method) (Reich, 2008; Perobelli et al., 2016).

The IOMs of 68 economic sectors were grouped and hierarchically organized into only five sectors by technological intensity, following the OECD's taxonomy of economic activities. This taxonomy is based on the sectors' investment intensity in research and development (R&D). Specifically, the technological intensity of sectors is determined by the fraction of private R&D investment relative to the gross value of sectoral production (Galindo-Rueda and Verger, 2016). This investment rate is used to classify sectors as having low, medium-low, medium, medium-high, and high technological intensity. A sixth category was created to include sectors that do not invest privately in R&D, such as the public sector (e.g., health, education, and security). The distribution of sectors in the IOMs, grouped by technological intensity, is detailed in the Appendix (Table A1).

The data on formal employment by level of education comes from the Relação Anual de Informações Sociais (RAIS-ME). From this database, information on formal employment by education level is extracted from approximately 670 economic activities at the class level. Subsequently, jobs in these economic activities are linked to the 68 sectors of input-output matrices, following the procedure provided by IBGE for matching the economic activities of CNAE 2.0 and SCN.7 Finally, employment is grouped by the level of human capital in sectors based on technological intensity.

For standardization, a nomenclature similar to that of sectors by technological intensity is used to distribute formal employment by level of human capital. Particularly, employment distributed in ten levels of education is grouped into only five levels of human capital: low (illiterate to 9th grade of elementary school), medium-low (complete elementary school and incomplete high school), medium (complete high school and incomplete higher education), medium-high (complete higher education), and high (master’s and doctoral degrees).

2.4 Limitations

Before presenting the results in the next section, it is important to highlight some limitations inherent to the empirical data and, consequently, to the adopted methodological approach. Firstly, the OECD taxonomy used to group sectors by technological intensity is based on R&D investments (Sarra et al., 2019). While the structure of the Brazilian economy differs from that of developed economies, Morceiro (2018) provides evidence that sectors with higher R&D investment in Brazil largely align with those identified in OECD countries, albeit with smaller-scale investments. The maintenance of this hierarchy suggests the suitability of the taxonomy for developing countries like Brazil. Nevertheless, this approach may still have the disadvantage of overlooking the characteristics and dynamics of investments in R&D in Brazilian sectors. Despite this limitation, the taxonomy remains valuable for grouping and classifying productive sectors based on technological intensity, especially in the absence of a national classification (Galindo-Rueda and Verger, 2016).

Regarding the labor market, it is important to note that this study focuses only on formal employment levels in Brazil, meaning it does not consider informal employment. This is due to some data limitations and the research scope. From the research perspective, the focus is on the endogenous factors of economic growth, such as the level of human capital of workers and the technological intensity of sectors. Using formal employment data, it is possible to map jobs by the level of human capital (educational level) and accurately group them into sectors by technological intensity over time. The same cannot be done with informal jobs, as these workers often do not appear in official statistics accurately and are strictly concentrated in low-productivity sectors. In contrast, the occupancy levels (occupied workforce) in the IOMs do not provide the characteristics of the workers, such as educational level or type of employment contract, thereby limiting their utility in this study.

A natural disadvantage of considering only formal employment in this analysis is the underestimation (or overestimation, in some cases) of results, as it considers only a fraction of the entire occupied workforce in the Brazilian economy - approximately 45% (Appendix A2). Despite this limitation, previous national studies have used formal employment by education level to conduct a structural decomposition analysis, such as Perobelli et al. (2016). Other studies found in the literature conducted similar analyses considering age (Carrascal Incera, 2017; Catelan et al., 2021), gender (Barba and Iraizoz, 2020), and workers' skills (Gregory et al., 2001). In other words, all of these studies conducted analyses of structural decomposition considering only part of the occupied workforce.

3 Results

This section presents the main empirical results of this study. The first part analyzes the evolution of formal employment multipliers by level of human capital. Subsequently, a structural decomposition analysis of employment is presented, both in aggregate and disaggregated form, based on the level of human capital and the technological intensity of the sectors.

3.1 Employment multiplier by level of human capital

The employment multiplier measures the ability of sectors to generate formal employment for each level of human capital in Brazil. Table 1 presents the ratio of sectoral employment multipliers by level of human capital. Workers with medium-low and medium levels of human capital have the highest proportions in sectoral employment multipliers. This indicates that fluctuations in final demand generate more jobs for workers with these qualification levels. Between 2010 and 2018, the proportions of medium, medium-high, and high levels of human capital increased in employment multipliers, while the proportion of lower levels gradually decreased in all sectors. Expanding the economy's demand generates jobs in sectors that require more skilled labor. The result also indicates a change in the labor market with a qualification bias, where more specialized workers find more job opportunities than less qualified workers (Acemoglu and Autor, 2011).

In 2018, the high-tech sector had an employment multiplier of 50.65% for workers with medium-level human capital. This indicates that 50.65% of job opportunities generated by the increase in final demand would go to workers with this level of qualification. High-level and medium-high human capital held 1.15% and 31.38% shares in the employment multiplier. The low-tech sector created most jobs for workers with middle- and lower-middle human capital, with 58.56% and 17.71% of the employment multiplier, respectively. As expected, other sectors, such as public administration, education, and health care, generate the most opportunities for highly skilled workers, at 2.63%. These sectors are the main employers of highly qualified human capital in the Brazilian formal labor market.

During the period of economic growth, there was a reduction in the employment multiplier in the low-tech sector for workers with low (19.3% to 14.84%) and medium-low (24.76% to 21.55%) qualifications. This reduction led to relative participation of 11.51% and 17.71% in the employment multiplier of the low-tech sector in 2018. On the other hand, the increase in demand positively affected the employment of workers with higher levels of human capital between 2010 and 2018. The medium-high level of human capital increased from 8.15% to 11.96%, while the high level increased from 0.13% to 0.26%.

Table 1:
Proportion of employment multipliers by level of human capital (%)

This trend of job creation for workers with higher levels of human capital is also observed in all sectors of technological intensity, indicating a correlation between a growth process of qualified job opportunities and economic and technological changes in Brazil. Furthermore, this change in the composition of the workforce, favoring those with higher levels of human capital, demonstrates the ability of the Brazilian productive structure to adapt and incorporate new technologies (Acemoglu and Restrepo, 2018; Perobelli et al., 2016).

3.2 Structural decomposition of formal employment

Table 2 presents the evolution of formal employment in sectors by technological intensity. Between 2010 and 2014, the Brazilian job market grew by 12.5% in formal jobs, from 44.1 million to 49.6 million, with an average employment growth rate of 3% per year. The positive highlights were in the low-tech sector, with an increase of 15.8% (3.57 million jobs). However, due to the economic contraction between 2014 and 2018, employment fell by 5.9%, from 49.6 million to 46.6 million. The “other sectors” were the only ones to show positive employment growth in the period, with an additional 111 thousand job positions. In this period, the low-tech sector lost 1.94 million jobs. Between 2010 and 2018, the low-tech (1.62 million), high-tech (8.84 thousand), and other (1.46 million) sectors had positive employment variations. In contrast, the medium-low and medium-high technology sectors lost 221 thousand and 205 thousand jobs, respectively. From 2010 to 2018, employment increased from 44.1 million to 46.6 million (5.8%), with an average growth rate of 0.75% per year.

Table 2:
Evolution of formal employment in sectors by technological intensity

Table 3 presents the structural decomposition of formal employment at aggregate levels for subperiods (growth, recession, and cycle). The analysis decomposes the variation of employment into three main components: labor intensity, technological change, and final demand. Labor intensity, denoted as labor change, measures the interaction between employment and production. A negative value for this factor indicates higher labor productivity, which occurs when production increases proportionally more than the growth of job opportunities. On the other hand, a positive value indicates a decline in labor productivity, suggesting that employment expansion did not keep pace with the increase in production. Technological change, on the other hand, arises from various elements that catalyze adjustments in the technical coefficients in the IOMs. This includes technological innovations, technical substitutions, and economies of scale. Lastly, the final demand encompasses changes in employment due to alterations in household and government consumption, exports, and gross fixed capital formation (investment). To simplify, we label this component as “capital” or “fixed capital”.

Between 2010 and 2014, employment growth was due to the effects of demand and technology, offset by labor intensity. Ceteris paribus, the increase in demand created 23.65 million jobs. Of this total, household consumption was the main driver of employment growth (12.29 million). Technology increased employment by 0.08 million, while labor intensity eliminated 18.24 million jobs, indicating an increase in productivity from this factor. Overall, employment increased by 11.1% (5.5 million).

From 2014 to 2018, the demand effect increased registered employment by 40.81 million, while the labor intensity effect decreased by 43.94 million. Technological change generated an increase of 0.19 million jobs, while the effect of exports doubled, creating 5.94 million jobs during the economic crisis. As a result, the total variation in employment was a 6.31% decrease, equivalent to a loss of 2.94 million jobs. From 2010 to 2018, the signs of the mentioned factors remained the same, although their magnitude increased. Technology, for example, increased employment by 0.36 million, while exports increased by 9.37 million and capital by 6.27 million. Final demand, led by household and government consumption, surpassed the use of labor in the Brazilian economy, resulting in an overall increase in employment of 5.5% (2.56 million).

Table 3:
Structural decomposition of employment at aggregate levels

Although technological unemployment is a reality in emerging markets, technological changes have increased employment in the Brazilian economy during the analyzed periods. Acemoglu and Restrepo (2018) argue that the workforce possesses a comparative advantage in new and more complex tasks that arise with technological progress and economic transformations. Supposing this comparative advantage is significant, and new tasks are created continuously, this would mean that employment can increase or remain stable in the long term, even with the introduction of innovations in the production system. It is important to emphasize that job creation is one of the main macroeconomic factors that stimulates economic development; therefore, it is essential to maintain a balance between employment in the economy and new technologies, as well as ensure that workers can adapt to economic changes (Souza Filho et al., 2021).

Figure 1 presents the structural decomposition of employment in sectors by technological intensity.8 The vertical axis shows the employment variation in percentage terms, while the horizontal axis represents the sectors by the technological intensity in subperiods. Each bar indicates the contribution of factors to the total variation of formal employment in sectors with different technological intensities. During the economic acceleration (2010-2014), there was a 13.6% increase in low-tech sector jobs in Brazil, equivalent to 3.5 million jobs. Household consumption increased employment by approximately 35%, followed by exports (7%) and government consumption (3%). While the contribution of capital to employment growth was not higher than 10%, work intensity acted in the opposite direction, reducing employment in sectors by over 40%. A similar pattern continued throughout the two subperiods (2014-2018 and 2010-2018) analyzed for the Brazilian economy. This shows the strong role of labor productivity growth in reducing formal employment opportunities in the country.

Figure 1
Structural decomposition of formal employment in sectors by technological intensity

During the economic recession (2014-2018), technical change created 320.2 thousand jobs in the low-tech sector, 72.7 thousand in the medium-high technology sector, and 142.7 thousand in other sectors. However, between 2010 and 2018, technological changes had mixed effects on job creation. Technical change generated jobs in the medium-high (36 thousand) and low (710 thousand) technology sectors and in other sectors (188 thousand) while eliminating jobs in the high (59.8 thousand), medium (258.4 thousand), and medium-low (255.3 thousand) technology sectors. Final demand components sustained formal employment growth between 2010 and 2018. Employment increased by 2.7 million in the Brazilian economy during this period.

Figure 2
Structural Change Index (SCI) for Brazil (2010-2018)

The Structural Change Index (SCI) shows instability in the participation of technology sectors in both employment and production. Between 2010 and 2018, the SCI had a value of 2.28, meaning that the sum of variations in the sectors' participation in employment was 2.28%. During economic acceleration and deceleration periods, the SCI for employment jumped from 1.49% to 1.96%. The SCI for production was 2.94% during economic growth and a lower value during the recession (1.66%). From 2010 to 2018, the SCI for production reached 3.63% (Table 3). The evolution of the SCI for employment and production over time confirms instability and reinforces the results observed for the subperiods (Figure 2).

Overall, the analysis of the SCI reveals structural changes in employment and production between 2010 and 2018. The results indicate that the observed instability may be directly linked to government economic interventions, particularly during the period from 2010 to 2014. These interventions encompassed strategies such as adjustments in monetary policy, regulatory interventions, and fiscal stimuli through investment policies like the Programa de Aceleração do Crescimento (PAC) and the Programa Minha Casa, Minha Vida (PMCMV). From 2014 to 2018, alterations in production and employment are likely attributed, at least in part, to the economic crisis (Magacho and Rocha, 2022; Bastos, 2017; Barbosa, 2017).

3.2.1 Decomposition of formal employment by level of human capital

This section presents the results of the structural decomposition of formal employment by the level of human capital during the periods of economic growth and recession in the 2010s. Firstly, Table 4 presents the evolution of formal employment by level of human capital. Employment opportunities for highly skilled workers have increased in Brazil. Between 2010 and 2018, the number of jobs requiring a high level of human capital rose from 213 thousand to 508 thousand, an extraordinary growth of 138.7%. Jobs for medium and medium-high human capital also increased from 20.26 million to 24.63 million (21.5%) and from 7.06 million to 10.22 million (44.8%), respectively. In contrast, employment requiring low to medium-low levels of human capital decreased from 9.29 million to 6.80 million (-26.8%) and from 7.24 million to 4.46 million (-38.3%), in this order.

Table 4:
Evolution of formal employment by level of human capital

The evidence suggests a trend toward the greater accumulation of qualified human capital. According to Perobelli et al. (2016), the persistence of this phenomenon is a key factor in the ability to incorporate technology into production processes, as well as in competitiveness and latent potential for economic growth.

Figure 3 presents the structural decomposition of employment for workers with low levels of human capital. The decomposition factors indicate the direction of percentage variation in formal employment concerning the level of human capital during growth and slowdown periods, as well as over the entire decade. Positive variations in the decomposition factors signify an increase in formal employment, while negative variations indicate job losses.

In absolute terms, around 1.1 million jobs for low-skilled workers were eliminated between 2010 and 2014 and 1.7 million between 2014 and 2018, resulting in a total reduction of 2.8 million job positions over the entire period. All sectors experienced a decrease in employment for low-educated workers, but the decline was most pronounced in the low-technology sector, which saw a reduction of 483.2 thousand jobs between 2010 and 2014 and 1.1 million between 2014 and 2018. Excluding the low technology sector and other sectors (public sector), the average reduction in employment for this group of workers was 75.9 thousand between 2010 and 2014, 100 thousand between 2014 and 2018, and 178 thousand between 2010 and 2018.

Figure 3
Structural decomposition of employment for the low level of human capital

The reduction in employment for less skilled workers is driven by the intensity of labor utilization, which decreases on average by 55% to 60% across sectors, indicating that sectoral production is growing more rapidly than the hiring rate. This result suggests that labor intensity is decreasing in less technology-intensive sectors. In the low-technology sector, for example, the reduction in employment is compensated by demand growth, with exports and household consumption dominating in each subperiod, representing an average of 7% and 30%, respectively. Technology also increased employment in the low-technology sector and other sectors (public sector) during all subperiods, although the effect was less than 5%. In contrast, in the medium-technology sector, technical changes reduced employment by up to 10%, as observed during the economic growth period.

Several aspects explain this phenomenon in Brazil during the 2010s. Firstly, technological advancement and automation significantly impacted low-tech sectors, such as agriculture and livestock. In these sectors, where repetitive and mechanical tasks can be easily automated, substituting labor with machines and automation has reduced the number of jobs for low-skilled workers. This process, together with the increase in international competitiveness and domestic consumption, has heightened the demand for higher-skilled workers at the expense of low-skilled workers. At the same time, over the last decade, the country has witnessed a widespread increase in the workforce's education level due to public policies aimed at decentralizing and facilitating access to higher education (Rada et al., 2019; Casqueiro et al., 2020).

The evidence suggests that these changes in the formal labor market are reducing the hiring of low-skilled workers due to their diminished adaptation to new economic factors, such as technological innovations and business cycles. This also reflects the growing need in technological sectors for a specialized workforce to increasingly engage in capital-intensive and technology-driven production systems (Acemoglu and Restrepo, 2018).

When analyzing the medium-low level of human capital, one can perceive that the labor factor was mainly responsible for reducing job opportunities (Figure 4). Between 2010 and 2018, the employment of this group suffered a negative variation of 2.5 million, with 283.1 thousand between 2010 and 2014 and 2.2 million between 2014 and 2018. Except for the low-technology sector, all other sectors showed reduced employment for this level of human capital during economic growth (2010-2014). The medium-low technology sector eliminated approximately 136.1 thousand jobs. On the other hand, the low-technology sector generated 51.3 thousand jobs, representing an increase of 0.9%, driven mainly by household consumption (30%), capital (10%), and exports (5%). However, job creation in this sector could have been even greater if it were not for the negative effect of the labor factor (labor intensity), which contributed to a 50% reduction in jobs.

This result highlights the decreasing demand for workers with lower levels of education in the Brazilian economy. Between 2014 and 2018, there was a significant contraction in employment for workers with medium-low levels of qualification, especially in the low (1.3 million) and medium-low (301 thousand) technology sectors. From 2010 to 2018, all sectors of the economy experienced reduced employment for workers with this qualification.

Figure 4
Structural decomposition of employment for the medium-low level of human capital

These findings are consistent with those of Perobelli et al. (2016), who concluded that the primary factor behind the decline in formal employment for the less skilled was the intensity of work following trade liberalization in Brazil. Furthermore, Sesso Filho et al. (2010) found results from 1991 to 2003 that do not differ significantly from those observed in the 2010s decade. They found that after trade liberalization in Brazil, labor productivity emerged as a key factor in reducing jobs in the agricultural and extractive sectors (low technology sectors). In contrast, the technological factor generated jobs in these sectors.

The transformations observed in the formal labor market context throughout the 2010s can be interpreted as a continuation of trends that had already been occurring since the 1990s and 2000s. However, it is essential to highlight that contemporary factors, such as economic cycles, educational policies, and technological innovations, have also influenced the change process that has occurred more recently in the Brazilian formal labor market.

Figure 5
Structural decomposition of employment for the medium level of human capital

Figure 5 highlights the employment decomposition for medium-skilled human capital. From 2010 to 2014, Brazil experienced an increase of 4.45 million jobs for this educational level, with all sectors showing employment growth during this period. The sectors that generated the most jobs were low (3.3 million), medium-low (435.8 thousand), and medium (122.1 thousand) technology. The driving forces for employment in low-tech sectors were primarily household consumption (30%), capital (10%), and exports (5%), offset by a decrease of the labor intensity (-32%). The medium-low-tech sector presented a pattern very similar to the previous sector. On the other hand, in the medium-tech sector, household consumption (31%), exports (18%), and capital (10%) increased employment for this group, while technology (-10%) and labor intensity (-25%) decreased. In absolute terms, exports generated 81 thousand jobs out of the 291.7 thousand created for this group due to demand variation in the medium technology sector. It is also worth noting the effect of capital, equal in magnitude to household consumption, in generating jobs in the high-tech sector, with 20% per each (105 thousand jobs). Although small, technical change generated medium-skilled human capital jobs in the low-tech (96.1 thousand) and high-tech (9.2 thousand) sectors. Additionally, other sectors saw an increase of 468.8 thousand job openings for this qualification level (8.5%), with growth driven by government consumption (33%).

During the economic recession (2014-2018), only the low-tech sector was able to create jobs for medium-skilled workers (179.4 thousand), thanks to the effects of technical change (1%) and demand (50%) outweighing the decrease in labor intensity (-50%). Excluding the low-tech sector, the remaining sectors experienced a reduction in employment for this group, with the medium-high (113 thousand) and high-tech (87.8 thousand) sectors being the most affected. The high-tech sector's job reduction was driven by both labor intensity (-55%) and technical change (-5%), while the main factor of unemployment in the medium-high-tech sector was changes in work requirements (-60%). Labor intensity decreased in these sectors during the economic crisis. From 2010 to 2018, the high- and medium-high-tech sectors eliminated approximately 100 thousand jobs, while other sectors generated about 4.46 million jobs for medium-skilled workers. Of this amount, the low-tech sector was responsible for 3.5 million jobs.

Figure 6
Structural decomposition of employment for the medium-high level of human capital

Figure 6 shows the employment decomposition for medium-high skilled workers. Employment in this group increased by 2.27 million from 2010 to 2014 and by 896.7 thousand from 2014 to 2018, with a total increase of 3.16 million jobs. All sectors had job growth in the first subperiod, with the public sector adding 1.1 million jobs (52%). The lower effect of work intensity was important for job growth, particularly in technology-intensive sectors. For example, work intensity had a negligible effect in medium-high technology sectors, while in medium-low technology sectors, it was less than 10%. Job growth in low- and medium-low-technology sectors was driven by household consumption (45% and 40%), capital (10% and 4%), and exports (7% and 11%). In medium- and medium-high-technology sectors, household consumption (35% and 45%), exports (20%), and capital (12% and 25%) were the main determinants of employment growth, while in high-technology sectors, household consumption (30%) and capital (22%).

The low-tech sector generated 661.5 thousand jobs for this type of human capital between 2010 and 2014, while the medium-low (219.1 thousand) and high-tech (114.4 thousand). In the low and medium-low technology sectors, employment growth was determined by household consumption (45% and 40%), capital (10% and 4%), and exports (7% and 11%). In the medium and medium-high technology sectors, household consumption (35% and 45%), exports (20%), and capital (12% and 25%) were the primary determinants of employment growth. Finally, in the high-tech sector, household consumption (30%) and capital (22%) were the main drivers.

During the economic downturn (2014-2018), job creation remained positive for workers with medium-high skill levels across all sectors. The low technology sector generated 330.6 thousand jobs, equivalent to a variation of 12%, whereas the medium-low and high technology sectors generated approximately 38.1 thousand (5.4%) and 32.7 thousand (7.85%) jobs, respectively. Demand accounted for approximately 55% of the change in employment in the low-tech sector, while it was 51% in the lower-middle and high-tech sectors. Household consumption and exports were the main drivers of job growth in low-tech sectors, but capital prevailed over export demand in sectors with higher levels.

From 2010 to 2018, the average change in employment for workers with a medium-high level of qualification was 32.2%. It is important to highlight that the other sectors (the public sector) created 1.66 million jobs for workers with this level of human capital during the decade, constituting 53% of all jobs created for this group. The government's consumption was one of the main drivers of employment growth for workers with this level of qualification, accounting for 40% of the positive variation.

These results are in line with the existing evidence for the Brazilian economy. For example, when examining the impacts of technology on the evolution of employment in the trade and service sectors throughout the 2000s in Brazil, Fiuza-Moura et al. (2016) identify a process of increasing labor productivity that correlates with the growing demand for skilled labor in both low- and high-tech sectors. Furthermore, factors related to final demand drove the growth of these opportunities. Similarly, Perobelli et al. (2016) emphasize that trade liberalization in Brazil has brought greater opportunities for professionals with high levels of human capital in low-tech and high-tech sectors, such as the chemical industry. Qualified labor has also been essential for the modernization and improvement of the provision of public services in the country. Notably, the Brazilian economy during the 2010s continued to follow this trend of greater integration of skilled labor into productive systems.

Despite the international crisis in 2007, Brazilian exports continued to contribute to job creation. Exports generated employment across all technological sectors. However, their impact on job creation was more pronounced in low-tech sectors. Throughout the 2010s, exports created approximately 950 thousand jobs for workers with a medium-high level of human capital. Overall, this increase in employment, with a greater emphasis on professional qualifications, is crucial for the growth of production, innovation, and the competitiveness of the national economy.

For workers with high levels of human capital (those with master’s and doctoral degrees), the structural decomposition presents a similar pattern to that of workers with medium-high qualifications for periods of economic acceleration and deceleration (Figure 7). Between 2010 and 2014, there was an increase in employment for this group of workers (134.8 thousand), with job opportunities concentrated in other sectors (111.7 thousand). The driver of employment growth in this sector is government consumption (72%). It is important to note that this workforce is mostly hired by the public sector and educational institutions, which is why it is concentrated in other sectors.

The other job positions created (23 thousand) are distributed among the low (10.6 thousand), medium-low (5.7 thousand), and high (4.7 thousand) technology sectors. Unlike what has been observed thus far, labor intensity positively contributed to employment growth for the more qualified workers in the medium-low (33%), medium (18%), and medium-high (8%) technology sectors, which indicates a loss of labor productivity. This means that in these sectors, the expansion of jobs did not keep pace with the increase in output. Despite this, the positive evolution of employment in technological sectors was primarily determined by final demand structure, especially household consumption.

Figure 7
Structural decomposition of employment for the high level of human capital

Despite the economic crisis between 2014 and 2018, 160.4 thousand jobs were created for workers with high human capital. The other sectors (public sector) created approximately 84% of job opportunities, with growth driven by government consumption (48%). Unlike the previous subperiod, work intensity reduced highly qualified employment in all sectors. Ceteris paribus, technical changes reduced jobs in the high-tech sector by 8.6 thousand for this group. From this perspective, the relative contribution of exports to job creation for these workers has ranged from 10% in the high-tech sector to 22% in the medium-high technology sector. Surprisingly, the effect of exports was equivalent to household consumption in the medium- and medium-high-tech sectors for this group of workers. However, in absolute terms, job creation is still insignificant.

In contrast, household consumption continued to be the main determinant of employment growth in the medium-low- and low-tech sectors. Between 2010 and 2018, approximately 295.2 thousand jobs were created, but their generation remained concentrated in other sectors (83%). Furthermore, the effect of the determinants remained similar to the previous period, with the structure of demand increasing job opportunities while technology and work intensity decreased.

The results indicate a low creation of employment opportunities for workers with master's and doctoral degrees in technological sectors, especially in non-public sectors. These results suggest that, despite technological advances and economic changes, the absorption of these highly qualified professionals still faces significant challenges in the formal Brazilian labor market, emphasizing the need for policies that create propitious environments for the development and utilization of the skills of this human capital in productive systems.

4 Discussion and conclusions

This paper explored endogenous factors of economic growth, such as the levels of human capital among workers and the technological intensity of sectors, to provide an analysis of the transformations that occurred in the formal labor market in Brazil during the 2010s. Sectoral multipliers reveal that job creation resulting from the expansion of final demand was directed toward workers with medium to high levels of human capital in all technological sectors. In contrast, employment for workers with lower levels of human capital has consistently decreased over the last decade. Consequently, the labor market has been shifting with a bias towards qualification, as even low-tech sectors are hiring workers with higher levels of education to enhance productivity and adapt to new economic factors. Overall, the economy is accumulating more qualified human capital in its production processes, thereby expanding its competitiveness and capacity to generate and absorb innovations to sustain increasing returns in production (Perobelli et al., 2016; Romer, 1990).

By decomposing the aggregate employment variation, it was possible to observe labor productivity gains in all sectors of the Brazilian economy during the periods of growth (2010-2014) and recession (2014-2018). The increasingly latent technological unemployment was not confirmed. In truth, technological changes modestly increased employment levels, suggesting that technical progress induced the creation of new activities and tasks that, in turn, translated into new employment opportunities (Acemoglu and Restrepo, 2018). This is the case for low- and medium-high-tech sectors, which progressively incorporate innovations to increase productivity and increasingly demand qualified labor. It is fundamental that this job creation movement in the economy remain balanced with new technologies and that workers become increasingly capable of adapting to economic changes (Souza Filho et al., 2021).

Although exports are essential to the Brazilian job market, in the most recent decade, job creation was dominated by household and government consumption expansion. The growth of exports oscillates between the second (2014-2018) and third (2010-2014) most important factors in the structure of final demand for the generation of formal jobs. However, according to Magacho and Rocha (2022), this engine lost strength between 2013 and 2016 due to the drop in commodity prices and, in recent times, has not been able to sustain the growth of production and employment at the levels observed in the early 2000s.

The consumption of households is the driving force behind employment growth in all technological sectors of the Brazilian economy due to its large domestic market. Government consumption is more important for job creation in low-tech sectors and in other sectors (public sector). Among the components of the final demand structure, the effect of gross fixed capital formation (investment) is greater than that of exports in generating jobs in the high-tech sector. In contrast, exports contribute more to job creation in low and medium-high technological intensity sectors. This result indicates that exports are increasingly contributing to employment growth in non-basic sectors of the economy. Maintaining this pattern is essential for decentralizing jobs in low-tech sectors and, consequently, for sustainable economic growth. These results are in line with other studies carried out on the Brazilian labor market, such as Sesso Filho et al. (2010), Perobelli et al. (2016), Fiuza-Moura et al. (2016), Catelan et al. (2021), and Acypreste (2022). Particularly, these studies show that labor productivity, although leading to increased unemployment in productive sectors, tends to be more than compensated by final demand.

The main conclusions regarding the decomposition of employment by the level of human capital in sectors by technological intensity are as outlined below. Firstly, low-skilled workers are the most affected by recent economic changes in Brazil, resulting in significant losses of formal jobs. On the other hand, employment opportunities have grown for higher-skilled workers in all technological sectors. These results indicate that the Brazilian labor market is following a trend toward hiring more specialized workers. The changes in the labor market are aligned with a transition to a more advanced economy, characterized by an increase in technological innovations and a demand for more skilled workers. Similar patterns were previously observed by Gregory et al. (2001) in the British labor market.

Secondly, the rise in unemployment, particularly for less skilled workers, is driven by a decrease in labor intensity (labor productivity growth). However, during economic growth (2010-2014), labor intensity had a small yet positive effect on job creation for more skilled workers in sectors with medium-high, medium, and medium-low technological intensity.

Thirdly, technological change has had small and mixed effects on formal employment variation. In low-tech sectors, technological changes are increasing the employment of more qualified workers, while in more technology-intensive sectors, there is a fluctuation between job creation and destruction. Although small labor-saving effects due to technical changes have also been observed, technological unemployment is more than offset by the growth in final demand.

Fourthly, the demand structure is also responsible for more than offsetting the job losses resulting from the reduction in labor intensity for more skilled workers in all technological sectors. For less skilled workers, the effect of labor intensity surpasses the final demand structure, especially in less technology-intensive sectors. Among the components of final demand, household consumption is the primary driver of employment growth at virtually all skill levels. In addition, government consumption, exports, and gross fixed capital formation (investment) also contributed to the increase in employment for workers with different levels of human capital.

As highlighted, changes in labor requisition and final demand have driven structural transformations in the formal labor market throughout the 2010s. The period of economic expansion was characterized by the hiring of more skilled workers by technological sectors. The increase in domestic consumption and favorable economic conditions, particularly between 2010 and 2014, helped to stimulate the formal job market. On the other hand, the economic crisis that extended from 2014 to 2018 had opposite effects. Overall, technological sectors are increasingly hiring qualified workers to meet the demands of an expanding consumer market.

In conclusion, the Brazilian economy experienced an increase of 2.56 million formal jobs between 2010 and 2018, with an average annual growth rate of 0.75%. To sustain this growth, it would be recommendable that the government stimulate the components of final demand, especially household consumption. This can be done, for example, through the resumption of public investments in infrastructure, such as those carried out through the Programa de Aceleração do Crescimento (PAC) and Minha Casa, Minha Vida (MCMV). In the short term, removing infrastructure investments from the government's spending cap may be necessary to make this initiative viable (Magacho and Rocha, 2022). This measure can help restore the pace of employment growth in Brazil.

Acknowledgements

This research was conducted with the support of the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). The authors thank CAPES and UFJF for the financial and institutional support granted to carry out this study. We also thank Professor Dr. Fernando Perobelli for his valuable comments and suggestions. Our thanks are extended to the Editor-in-Chief of Nova Economia, Dr. Alexandre Cunha, and to the two anonymous reviewers, whose contributions were fundamental to improving this paper.

References

  • ACEMOGLU, D. Technical change, inequality, and the labor market. Journal of economic literature, v. 40, n. 1, p. 7-72, 2002.
  • ACEMOGLU, D.; AUTOR, D. Skills, tasks and technologies: Implications for employment and earnings. In: Handbook of labor economics. Elsevier, 2011. p. 1043-1171.
  • ACEMOGLU, D.; AUTOR, D.; HAZELL, J.; RESTREPO, P. Artificial intelligence and jobs: Evidence from online vacancies. Journal of Labor Economics, v. 40, n. S1, p. S293-S340, 2022.
  • ACEMOGLU, D.; RESTREPO, P. Automation and new tasks: How technology displaces and reinstates labor. Journal of economic perspectives, v. 33, n. 2, p. 3-30, 2019.
  • ACEMOGLU, D.; RESTREPO, P. The race between man and machine: Implications of technology for growth, factor shares, and employment. American economic review, v. 108, n. 6, p. 1488-1542, 2018.
  • ACYPRESTE, R. Employment, technical change and growth in Brazil: a conclusion from Input-Output Matrix. Brazilian Journal of Political Economy, v. 42, p. 460-480, 2022.
  • AUTOR, D. H.; LEVY, F.; MURNANE, R. J. The skill content of recent technological change: An empirical exploration. The Quarterly journal of economics, v. 118, n. 4, p. 1279-1333, 2003.
  • BARBA, I.; IRAIZOZ, B. Effect of the great crisis on sectoral female employment in Europe: A structural decomposition analysis. Economies, v. 8, n. 3, p. 64, 2020.
  • BARBOSA, F. D. H. A crise econômica de 2014/2017. Estudos Avançados, v. 31, n. 89, p. 51-60, 2017.
  • BASTOS, P. P. Z. Ascensão e crise do governo Dilma Rousseff e o golpe de 2016: poder estrutural, contradição e ideologia. Revista de Economia Contemporânea, v. 21, p. e172129, 2017.
  • CARRASCAL INCERA, A. Drivers of change in the European youth employment: a comparative structural decomposition analysis. Economic Systems Research, v. 29, n. 4, p. 463-485, 2017.
  • CASQUEIRO, M. L.; IRFFI, G.; SILVA, C. C. A expansão das Universidades Federais e os seus efeitos de curto prazo sobre os Indicadores Municipais. Revista da Avaliação da Educação Superior (Campinas), v. 25, p. 155-177, 2020.
  • CATELAN, D. W.; SOUZA, K. B.; ALMEIDA VALE, V. Determinantes da Variação do Emprego dos Jovens Brasileiros: Uma Análise de Decomposição Estrutural para 2002-2015. Revista Brasileira de Estudos Regionais e Urbanos, v. 15, n. 2, p. 310-344, 2021.
  • COMPAGNUCCI, F.; GENTILI, A.; VALENTINI, E.; GALLEGATI, M. Have jobs and wages stopped rising? Productivity and structural change in advanced countries. Structural Change and Economic Dynamics, v. 56, p. 412-430, 2021.
  • CUNHA, A. M.; LÉLIS, M. T. C.; LINCK, P. Flutuações no nível de atividade e os ciclos de preços de commodities: evidências para o Brasil. Brazilian Journal of Political Economy, v. 41, p. 466-486, 2021.
  • DIETZENBACHER, E.; LOS, B. Structural decomposition techniques: sense and sensitivity. Economic Systems Research, v. 10, n. 4, p. 307-324, 1998.
  • DOAN, H. T. T.; LONG, T. Q. Technical change, exports, and employment growth in China: A structural decomposition analysis. Asian Economic Papers, v. 18, n. 2, p. 28-46, 2019.
  • FIUZA-MOURA, F. K.; NAKATANI-MACEDO, C. D.; CAMARA, M. R. G.; SESSO FILHO, U. A. Tecnologia e emprego nos setores comércio e de serviços no Brasil entre 2000 e 2009. Revista Brasileira de Inovação, v. 15, n. 1, p. 87-112, 2016.
  • GALINDO-RUEDA, F.; VERGER, F. OECD taxonomy of economic activities based on R&D intensity. OECD Science, Technology and Industry Working Papers. Paris: OECD Publishing, 2016.
  • GLOMM, G.; RAVIKUMAR, B. Public versus private investment in human capital: endogenous growth and income inequality. Journal of political economy, v. 100, n. 4, p. 818-834, 1992.
  • GREGORY, M.; ZISSIMOS, B.; GREENHALGH, C. Jobs for the skilled: how technology, trade and domestic demand changed the structure of UK employment, 1979-90. Oxford Economic Papers, v. 53, n. 1, p. 20-46, 2001.
  • GUILHOTO, J. J. M.; SESSO FILHO, U. A. Estimation of the input-output matrix using preliminary data from national accounts: application and analysis of economic indicators for Brazil in 2005. Economy & Technology, v. 6, p. 15-16, 2010.
  • HAUKNES, J.; KNELL, M. Embodied knowledge and sectoral linkages: An input-output approach to the interaction of high-and low-tech industries. Research Policy, v. 38, n. 3, p. 459-469, 2009.
  • LIMA, A. F. R.; RIBEIRO, F. G.; TEIXEIRA, G. S. Job qualification programs for young apprentices: an analysis in the Brazilian Midwest. Education+Training, v. 62, n. 6, p. 677-692, 2020.
  • LISE, J.; POSTEL-VINAY, F. Multidimensional skills, sorting, and human capital accumulation. American Economic Review, v. 110, n. 8, p. 2328-2376, 2020.
  • MACEDO, F. C.; PORTO, L. R. Evolução Regional do Mercado de Trabalho no Brasil (2000-2018): apontamentos para a política nacional de desenvolvimento regional (PNDR). Texto para Discussão (No. 2652). Brasília: IPEA, 2021.
  • MADARIAGA, R. Factors driving sectoral and occupational employment changes during the Spanish boom (1995-2005). Economic Systems Research, v. 30, n. 3, p. 400-421, 2018.
  • MAGACHO, G. R.; ROCHA, I. L. Demand-led growth decomposition: an empirical investigation of the Brazilian slowdown in the 2010s. Cambridge Journal of Economics, v. 46, n. 2, p. 371-390, 2022.
  • MILLER, Ronald E.; BLAIR, Peter D. Input-output analysis: foundations and extensions. Cambridge university press, 2009.
  • MORCEIRO, P. C. A indústria brasileira no limiar do século XXI: uma análise da sua evolução estrutural, comercial e tecnológica. 2018. Tese de Doutorado. Universidade de São Paulo.
  • NASSIF, A. An analysis of Brazil’s economic situation: 2014-2017, the short-term outlook and policy alternatives. Brazilian Keynesian Review, v. 3, n. 1, p. 95-108, 2017.
  • PEROBELLI, F. S.; BASTOS, S. Q. A.; PEREIRA, M. Z. Decomposição estrutural do emprego por grau de instrução: uma análise de insumo-produto para o período pós-abertura (1990 a 2005). Nova Economia, v. 26, p. 909-942, 2016.
  • RADA, N.; HELFAND, S.; MAGALHÃES, M. Agricultural productivity growth in Brazil: Large and small farms excel. Food policy, v. 84, p. 176-185, 2019.
  • REICH, U. Additivity of deflated input-output tables in national accounts. Economic Systems Research, v. 20, n. 4, p. 415-428, 2008.
  • ROMER, P. M. Endogenous technological change. Journal of political Economy, v. 98, n. 5, Part 2, p. S71-S102, 1990.
  • ROSE, A.; CASLER, S. Input-output structural decomposition analysis: a critical appraisal. Economic systems research, v. 8, n. 1, p. 33-62, 1996.
  • SARRA, A.; BERARDINO, C.; QUAGLIONE, D. Deindustrialization and the technological intensity of manufacturing subsystems in the European Union. Economia Politica, v. 36, n. 1, p. 205-243, 2019.
  • SESSO FILHO, U. A.; RODRIGUES, R. L.; MORETTO, A. C.; BRENE, P. R. A.; LOPES, R. L. Decomposição estrutural da variação do emprego no Brasil, 1991-2003. Economia Aplicada, v. 14, p. 99-123, 2010.
  • SOUSA FILHO, J. F.; SANTOS, G. F.; SANTANA RIBEIRO, L. C. Encadeamentos produtivos e multiplicadores macrossetoriais no Brasil após a abertura comercial. Revista de Economia, v. 42, n. 79, p. 670-698, 2021.
  • TIN, P. B. A Decomposition Analysis for Labour Demand: Evidence from Malaysian Manufacturing Sector. WSEAS Transactions on Business and Economics, 11, 32-41, 2014.
  • Supplementary Material
    Supplementary Material (Excel Sheet) contains additional results and information of secondary interest to readers and policymakers. Access: https://bit.ly/3VDFS8O
  • JEL Codes:
    C67, J20, O30, R10.
  • Códigos JEL:
    C67, J20, O30, R10.
  • 1
    Specifically, it involves economic distortions caused by broad government intervention, including implementing interest rate cuts, price controls, targeted investments, and subsidies. Furthermore, the abrupt slowdown in the global economy and the economic stagnation of major trading partners such as China and Europe further exacerbated the economic crisis in Brazil (Magacho and Rocha, 2022; Barbosa, 2017).
  • 2
    A precise definition is presented in the data section.
  • 3
    A precise definition is presented in the data section.
  • 4
    The Supplementary Material includes employment and production multipliers derived from open and closed input-output models. For a comprehensive understanding of the modeling, see Miller and Blair (2009). In this paper, our analysis will focus exclusively on simple employment multipliers.
  • 5
    We opted to use the polar mean due to its broad acceptance as a weighting method in structural decomposition analysis (e.g., Perobelli et al., 2016; Sesso Filho et al., 2010). Although other weighting methods can be used, Dietzenbacher and Los (1998) recommend using the polar mean as a more robust alternative for structural decomposition analysis.
  • 6
    The Núcleo de Economia Regional e Urbana da Universidade de São Paulo (NEREUS) provided the input-output matrices (IOMs) used in this paper.
  • 7
    For correspondence between activities in the Sistema Nacional de Contas (SNC) and the Classificação Nacional de Atividades Econômicas (CNAE 2.0), we followed the documentation provided on the IBGE website (https://concla.ibge.gov.br/documentacao/documentacao-cnae-2-0.html).
  • 8
    The detailed results of the structural decomposition of employment into sectors by technological intensity are presented in Appendix A3 (Table A2).

APPENDIX

A1 Sectors of the IOM grouped by technological intensity
Table A1:
Industries grouped into sectors by technological intensity
A2 Formal employment and occupied workforce in Brazil

Figure A1
Formal employment and occupied workforce in Brazil

Figure A2
Average share of formal employment in the occupied workforce

A3 Structural decomposition of formal employment in Brazil
Table A2:
Decomposition of employment into sectors by technological intensity
Table A3:
Decomposition of employment for the low level of human capital
Table A4:
Decomposition of employment for the medium-low level of human capital
Table A5:
Decomposition of employment for the medium level of human capital
Table A6:
Decomposition of employment for the medium-high level of human capital
Table A7:
Decomposition of employment for the high level of human capital

Publication Dates

  • Publication in this collection
    22 Nov 2024
  • Date of issue
    Apr-Jun 2024

History

  • Received
    14 July 2023
  • Accepted
    12 Mar 2024
location_on
Nova Economia FACE-UFMG, Av. Antônio Carlos, 6627, Belo Horizonte, MG, 31270-901, Tel.: +55 31 3409 7070, Fax: +55 31 3409 7062 - Belo Horizonte - MG - Brazil
E-mail: ne@face.ufmg.br
rss_feed Acompanhe os números deste periódico no seu leitor de RSS
Acessibilidade / Reportar erro