MANUFACTURING DEVELO PMENT? FLEXIBLE WORKERS AND THE FEMINIZATION OF SKIL LED JOBS IN THE MANA US INDUSTRIAL DISTRI CT By SKYLER MARK SIMNITT A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2014
Â© 2014 Skyler Mark Simnitt
4 ACKNOWLEDGMENTS I want to thank all the research participants in Manaus, who shared with me their stories and provided insight into the phenomena of flexible workers and the gendered division of labor in the Manaus Indust rial District. I also want to give a special thanks to the Almas family for their generous hospitality and support while I was in Manaus, allowing me to make the most of my time there. I thank the Center for Latin American Studies at the University of Flor ida who provided me with a Tinker Travel Grant to conduct fieldwork in the summer of 2013, and Dr. Charles Wood who provided me with the necessary data sets to carry out the quantitative portion of this research. I also want to thank my thesis committee me mbers Dr. Brian Gendreau and Dr. Connor Mullally for their invaluable feedback and support during the writing process. Above all I give a very special thanks to Dr. Carmen Deere, for her patience, support and guidance throughout every step of this project .
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBR EVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 Views of Development ................................ ................................ ............................ 19 Gender and Development ................................ ................................ ....................... 25 The Feminization of Labor ................................ ................................ ................ 26 The Defeminization of Labor ................................ ................................ ............ 28 Conclusion ................................ ................................ ................................ .............. 30 2 THE MANAUS FREE TRADE ZONE: ORIGINS AND CHALLENGES ................... 32 ................................ ................................ ..................... 32 The Amazon: A History of Economic Cycles and Fiscal Incentives .................. 33 Creation of the MFZ ................................ ................................ ......................... 36 The Free Trade Zone and Regional Development ................................ ........... 38 Macro Economic Policies and the MFZ ................................ ............................ 40 The MFZ as an Employer ................................ ................................ ................. 41 Assemb ly Workers ................................ ................................ ........................... 43 Female Workers ................................ ................................ ............................... 45 Restructuring ................................ ................................ ................................ .......... 47 Fordism as the Traditional Industrial Model ................................ ...................... 47 Flexible accumulation and Toyotism ................................ ................................ 48 Opening of the Brazilian Economy ................................ ................................ ... 50 Restructuring and Precarious Employment in Manaus ................................ ..... 50 Labor Unions and the MFZ ................................ ................................ ..................... 54 Conclusion ................................ ................................ ................................ .............. 57 3 CHANGING TRENDS WITH REGARDS TO GENDER OF THE WORKFORCE ... 61 Hypothesis ................................ ................................ ................................ .............. 63 Methods ................................ ................................ ................................ .................. 64 Terminology ................................ ................................ ................................ ...... 65 The Data ................................ ................................ ................................ ........... 65
6 Selecting the Population of Study ................................ ................................ ..... 66 Tools of Analysis ................................ ................................ .............................. 68 Recoding ................................ ................................ ................................ .......... 69 Means Comparison ................................ ................................ .......................... 70 Findings ................................ ................................ ................................ .................. 71 Economically Active Population ................................ ................................ ........ 71 Educational Attainment ................................ ................................ .................... 72 The Manufacturing Sector ................................ ................................ ................ 74 Gendered Distribution of the Workforce in the Manaus Industrial District ........ 75 Skil led Jobs ................................ ................................ ................................ ...... 77 Differences in Earnings between Men and Women ................................ .......... 78 Conclusion ................................ ................................ ................................ .............. 80 4 FLEXIBLE WORKERS AND GENDER ................................ ................................ ... 96 Hypotheses Regarding Flexible Workers ................................ ................................ 97 Hypotheses Regarding Gender ................................ ................................ .............. 98 Organization ................................ ................................ ................................ ........... 99 Methods ................................ ................................ ................................ .................. 99 Terminology ................................ ................................ ................................ ...... 99 Sampling Method ................................ ................................ ........................... 101 Eligibility of Participants ................................ ................................ .................. 102 Locating Participants ................................ ................................ ...................... 103 Key In formants ................................ ................................ ............................... 104 Semi Structured Interviews ................................ ................................ ............ 105 Coding ................................ ................................ ................................ ............ 106 Sample Characteristics ................................ ................................ ................... 107 Find ings ................................ ................................ ................................ ................ 109 Flexible vs. Fully Contracted Workers ................................ ............................ 109 Differences in education ................................ ................................ .......... 109 Differences in pay and benefits ................................ ................................ 114 Differences in treatment ................................ ................................ ........... 118 Professional goals ................................ ................................ .................... 120 Gender Analysis ................................ ................................ ............................. 120 Educational attainment ................................ ................................ ............ 122 Professional qualifications ................................ ................................ ........ 122 Women in skilled positions ................................ ................................ ....... 123 Gendered division of assembly processes ................................ ............... 126 Sexual harassment and exploitation ................................ ........................ 128 Future plans by gender ................................ ................................ ............ 130 How Workers View the Manaus Free Trade Zone ................................ .......... 131 The Household ................................ ................................ ............................... 133 Marital status, living arrangements ................................ .......................... 133 Budget and finances ................................ ................................ ................ 135 Severance Pay ................................ ................................ ......................... 138 Conclusion ................................ ................................ ................................ ............ 139
7 5 FINAL CONSIDERATIONS ................................ ................................ .................. 181 LIST OF REFERENCES ................................ ................................ ............................. 189 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 198
8 LIST OF TABLES Table page 2 1 Mean earnings by manufacturing subsector a comparison of MID and SÃ£o Paulo  (currency values BRL$) ................................ ................................ . 60 2 2 Unemployment rates of Brazil's 20 most populous cities (2010) ........................ 60 3 1 Change in workforce at the Manaus Industrial District 1998 2013 (annual averages) ................................ ................................ ................................ ........... 85 3 2 Total population of the state of Amazonas and Manaus and % of population that is economically active ................................ ................................ .................. 86 3 3 Incidence educational attainment of the economically active population of Manaus ................................ ................................ ................................ ............... 87 3 4 Incidence educational attainment of manufacturing workforce, by gender ......... 88 3 5 Gendered composition of MID workforce, disaggregated by subsector (1970 2010) ................................ ................................ ................................ ............... 89 3 6 Distribution of skilled MID workforce by gender, disaggregated by subsector (1970 2010) ................................ ................................ ................................ ........ 90 3 7 Monthly earnings of MID workforce, disaggregated by gender [1970 2010] (currency values BRL$, adjusted to 2010) ................................ .......................... 91 3 8 Share of MID Workforce in skilled positions, disaggregated by gender subsector ................................ ................................ ................................ ............ 92 4 1 Respondents' age groups, disaggregated by gender and employee type ........ 144 4 2 identified color/race ................................ ............................. 145 4 3 Skilled vs. unskilled workers, disaggregated by gender and employee type .... 146 4 4 High school graduates among respondents, disaggregated by gender and employee type ................................ ................................ ................................ .. 147 4 5 Number of training certificates, disaggregated by gender and employee type . 148 4 6 Educational attainment tertiary schools, disaggregated by employee type and gender ................................ ................................ ................................ .............. 149 4 7 Living arrangement, disaggregated by gender and employee type .................. 150
9 4 8 Number of children among participants, disaggregated by gender and living arrangement ................................ ................................ ................................ ..... 151 4 9 Ages of workforce (electronics subsector only), disaggregated by gender Manaus 2010 ................................ ................................ ................................ .... 152 4 10 Composition of the electronics manufacturing workforce by marriage status, disaggregated by gender ................................ ................................ .................. 153 4 11 Educational attainment tertiary schools, disaggregated by gender and employee type ................................ ................................ ................................ .. 154 4 12 Educational attainment tertiary schools, disaggregated by gender and employee (skilled workers excluded from analysis) ................................ .......... 155 4 13 Number of training certificates, disaggregated by gender and employee type . 156 4 14 Number of training certificates (only unskilled workers considered), disaggregated by gender and employee type ................................ ................... 157 4 15 Principal benefits participants reported receiving from their employers, disaggregated by employee type ................................ ................................ ...... 158 4 16 Mean base earnings, disaggregated by employee type (currency BRL$) ........ 159 4 17 Mean base earnings of unskilled workers, disaggregated by employee type (currency BRL$) ................................ ................................ ............................... 160 4 18 Participants' assessment of the Manaus Free Trade Zone Main themes emerging in responses, disaggregated by employee type and gender ............. 161 4 19 Reason for leaving last job in the MID, responses disaggregated by gender and employee type ................................ ................................ ........................... 162 4 20 Perceived difference in treatment of flexible workers compared to fully contracted workers, responses disaggregated by gender and employee type (multi answers allowed) ................................ ................................ .................... 163 4 21 Future plans with regards to employment/profession coded by employee type and gender (multi answers allowed) ................................ ................................ . 164 4 22 High school graduates among respondents, disaggregated by gender and employee type ................................ ................................ ................................ .. 165 4 23 Number of training certificates, disaggregated by gender (both skilled and unskilled workers included) ................................ ................................ ............... 166 4 24 Are there more males or females employed in your work area? Responses disaggregated by gender ................................ ................................ .................. 167
10 4 25 Sex of overseer, disaggregated by respondents' gender ................................ .. 168 4 26 Content analysis, gendered themes most addressed in interview, disaggregated by individual respondents (multi answers allowed) 1 .................. 169 4 27 Future plans with regards to employment/profession coded by gender and employee category (multi answers allowed) ................................ ..................... 170 4 28 Marital status of respondents, disaggregated by gender ................................ .. 171 4 29 Living arrangement, disaggregated by marital status and gender .................... 172 4 30 Who is primarily resp onsible for paying the given expense? Disaggregated by gender and living arrangements ................................ ................................ .. 173 4 31 Number of children among participants, disaggregated by gender and living arrangement ................................ ................................ ................................ ..... 175 4 32 Participants with children living in the home, disaggregate d by gender and living arrangement ................................ ................................ ............................ 176 4 33 After you have paid all household bills and expenses,* how much of your earnings are left for your personal use?* ................................ .......................... 177 4 34 Average gross monthly earnings workers with children in the home, disaggregated by gender ( currency BRL$) ................................ ....................... 178 4 35 Average gross monthly earnings unskilled workers with children in the home, disaggregated by gender (currency BRL$) ................................ ....................... 179 4 36 How did you use your severance pay? Disaggregated by gender and employee type (only one answer considered) ................................ .................. 180
11 LIST OF FIGURES Figure page 3 1 Historical variation of the workforce in the Manaus Industrial District ................. 84 3 2 Composition of the MID workforce by skilled and unskilled jobs ......................... 93 3 3 Wage gap between male and female unskilled MID workers ............................. 94 3 4 Wage gap between male and female skilled MID workers (1970 2010) ............. 95
12 LIST OF ABBREVIATIONS M FZ Manaus Free Trade Zone M ID Manaus Industrial District Suframa The Superintendence of the Manaus Free Trade Zone ( SuperintendÃªncia da Zona Franca De Manaus )
13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts MANUFACTURING DEVELO PMENT? FLEXIBLE WORKERS AND THE FEMINIZATION OF SKILLED JOBS IN THE MANAUS INDUSTRIAL DISTRICT By Skyler Mark Simnitt August 2014 Chair : Carmen Diana Deere Major: Latin American Studies The themes of gender, restructuring, and wage employment converge in the through state led industrialization in the Americas. This research addresses the participation of women in It includes an original analysis of Brazilian National Census data, and data collected through qualitative interviews. Findings include that the manufacturing workforce in Manaus significantly defeminized with the liberalization of the Brazilian economy after 1990, with the ratio of females relative to males employed in the sector decreasing by positions actually in creased after 1991. This is further substantiated by interviews with increased in recent years. Although 10% of the workforce in the Manaus Industrial District currently con sists of flexible workers (individuals employed under temporary and subcontracts) this important class of employees is rarely given attention in the literature. This relatively
14 marginalized group of factory workers is analyzed here through interview data. Findings indicate that there is no discern i ble difference between flexible workers fully contracted attainment. Also, both classes of workers reported similar levels of job satisfaction, and anxiety about job stability even though flexible workers generally receive fewer benefits fully contracted time contracts).
15 CHAPTER 1 INTRODUCTION On the 30 th of September, 1968, four miles from downtown Manaus, scores of curious onlookers huddled under makeshift palm frond pavilions to mark a historic occasion. Presiding at the event were Governor Danilo Matos of the state of Amazonas and industrial district Superintendent Floriano Pacheco. The dirt plaza was mostly unadorned, save the aforementioned structures, and a fifty foot banner, stretching the length of the locale, which read, INDUSTRIAL DISTRICT: Point of Redemption for the We stern Amazon. 1 The crowd cheered as the symbolic cornerstone was lowered into its development to Brazil the most conspicuous element of the Manaus Free Trade Zone researchers have written interest is rooted in the perceived exoticism of the location, as it sits in the midst of the largest tropical forest and river system in the world, and is located thousands of qualifying busi nesses within its limits receive the most generous fiscal incentives in the country. As a major employer, the industrial district has transformed the city of Manaus from a midsized port community into a major urban center and served as a pole of attraction for migrants from surrounding and distant regions of the country. 1 . (SerÃ¡fico M. , 2011 : 23 )
16 Although Manaus remains less developed than other parts of Brazil, the project generates revenue for the state government, and provides thousands of industrial jobs, with a spillover effec t into other economic sectors (Ferreira, 2003) . Individuals from rural areas who have moved to Manaus in search of employment, have experienced an increased standard of living as they have gained access to basic services such a s healthcare and education. Ultimately the free trade zone has converted Manaus into one of the wealthiest cities in the Brazilian North region. For students and scholars of development, the Manaus Free Trade Zone provides a particularly fascinating contex t to study the effects of state led industrialization on a clearly delimited, and remote urban center. In its more than 47 years of history, the project has survived major changes in the national and global economies, at times reflecting external shocks wi th reductions in the workforce, geographic isolation from the rest of the country, and its primary dependence on the domestic consumer market. For these and other reasons, the Manaus Free Trade is markedly different than other free trade areas addressed in the literature, which serve external markets. I first became acquainted with the Manaus Free Trade Zone while traveling in prominently in the lives of many people I spoke with, whether they themselves or their relatives had worked there. Having spent several weeks previously in the neighboring state of Acre, where virtually everyone I met either worked in small scale commerce or
17 I understood little about tax incentives and state led development, however, the peculiarity of the Manus Industrial Distri ct in terms of its location and scope left an impression on me, evidently later influencing the course of my graduate studies. What follows is a presentation of original research on changing trends in the workforce in the Manaus Industry District (MID) w ith regards to gender and workers under flexible contracts. Through the analysis of Brazilian national census data I seek to determine how the share of jobs occupied by women in the industrial district has changed over the last four decades. Th rough the an alysis of interview data with factory workers in Manaus, I assess differences between fully contracted factory workers and their colleagues employed under flexible contracts. My findings contribute to the literature on gender in in the Manaus Free Trade Zo ne (MFZ) since few studies over a period of decades. Additionally, my findings regarding workers in the MID under flexible contracts are beneficial to understanding labor flexi bility in the region, since there is a lacuna in the literature with regards to these types of workers. One of the main questions I address in my research centers on how the share of industrial jobs occupied by women in the Manaus Industrial District h as changed since 1970, particularly since the opening of the Brazilian economy in 1990. I approached manufacturing sector generally increased between 1970 and 2010. Desp ite the literature arguing that the maturation of industrial processes and technological advances favor male over female workers, I assumed that the superior educational attainment levels of women in Brazil, along with the historic presence of females in u nion
18 leadership in Manaus, would have counteracted such trends observed elsewhere. Another important question I address in my research centers on the use of flexible labor in the MID. I seek to answer how the workers employed under temporary and subcontrac ts (flexible workers) differ from their fully contracted counterparts in terms of pay, benefits, and other qualifications. As put forth in the literature, increased labor flexibility aids businesses in their accumulation of capital as they are better able to adjust production to meet fluctuations in demand for their products. However, as flexible labor models often include a preference for short term over long term contracts, many workers experience less stable employment with fewer benefits thus labor fle xibilization (Druck, 2011) . Through my research I seek to contribute to one of the fundamental questions posed by critics of the Manaus Free Trade Zone, namely, to what ext ent is the Manaus Free Trade Zone providing living wage jobs to the local population? The organization of this manuscript is as follows: The remainder of Chapter 1 is devoted to a brief review of the literature on development and gender in the context of Latin America and Brazil. The Chapter 2 is devoted to a review of the literature on the Manaus Free Trade Zone and the manufacturing workforce in the industrial district. Also included in Chapter 2 fically as it pertains to factory labor. Chapter 3 presents the quantitative portion of the research based on cross sectional data, and relates the main findings regarding the gender composition of th e workforce in Manaus. Chapter 4 describes the qualitat ive research I carried out in Manaus, and my analysis and main findings regarding flexible workers.
19 Chapter 5 is devoted to a brief summary of the issues, conclusions, and main findings of this research. Views of Development While the United Nations has (LDCS, 2014) . Implicit in such terms is the not become modernized or superior to what they once were. industrialization, economic growth, and per capita GDP, since the 1970s scholars and policy makers have humanized the concept. Among the most influential of these theorists is Amartya Sen who successfully argued, that an essential component of to live the lives they choose (Kingstone, 2011: 15). Sen was instrumental in the creation of the Human Development Index (HDI), a composite index that aggregates elements such as education levels, life expectancy, and income distribution into an ind ex value that can be used to quantify the development level of a given nation, and rank countries for purposes of comparison. In more recent years policy makers and development theorists increasingly consider human and legal rights as requisites for devel opment. Thus, particularly in the
20 regarding the labor market and job availability are additionally pertinent to discussions of development (Campo s, 2011) . 2 Currently Brazil ranks 85 out of 186 nations in terms of HDI, placing it among the upper half of nations in terms of human development but still significantly below most OECD member countries. 3 , 4 Brazil has also faired relatively well in terms of per capita GDP, which as of 2012 is $11,630 annually. 5 According to the World Bank this places Brazil firmly within the tier of upper middle income countries. Despite moderate rankings in the aforementioned indicators, Brazil suffers from pronou on income inequality, Brazil ranks 142, towards the tail end. These rankings include the GINI coefficient, an index that assesses the relative income inequality within a natio n (Volscho, 2008) . Brazilian government officials and policy makers are cognizant of the level inequality and have implemented programs aimed at ameliorating the problem through the redistribution of income. Among the most rese arched and highly acclaimed of these programs is the Bolsa FamÃlia , a conditional cash transfer program that provides supplemental income to families in poverty and aims to reduce intergenerational inequality. Although the Bolsa FamÃlia program has succeed ed at 2 . The right to work is among the basic human rights listed in the UN Universal Declaration of Human Rights (1948). While a lack of civil liberties does not necessarily preclude development, scholars of development and democracy have consistently pointed out the relationship between the two (Diamond L, 1994). 3 . Human Development Index, data available United Nations Development Programme: Human Development Reports. < http://hdr .undp.org/en/data >. 4 . The acronym OECD refers to the Organization for Economic Co Operation and Development, an organization consisting of member countries who have achieved requisite levels of prosperity and development. Included among the members are th e United States, the United Kingdom, Japan, several Nordic countries, and others. 5 . Data from the World Bank website, list of nations by comparative GDP.
21 lifting thousands of families out of abject poverty, more drastic and long term solutions are needed if Brazil is to achieve income equality levels similar to OECD member countries. The general population is acutely aware of inequality, corruption, a nd other social problems in their country. Last year this led thousands of students in Brazil to engage in demonstrations and street protests demanding social assistance, fiscal transparency, and harsher penalties for politicians who engage in corrupt prac tices (Nascimento, 2013) economy, in contrast to its failings in terms of human development, frequently contributes to an inferiority complex among many educated Brazilians (M ariotti, 2014) . The reasons Brazil, despite its abundant natural resources and geographic advantages, has failed to develop to the level of OECD nations are complex and varied. Authors have filled volumes discussing Brazilian development in historica l and contemporary contexts. While a comprehensive explanation of the factors that developed nation are beyond the scope of this paper, it suffices to say that Brazil has followed a similar devel opment path to other Latin American nations, including dependence on an export model neither conducive to long term growth nor the expansion of sectors that have proven critical in the development paths of other countries. Currently Brazil is the most industrialized nation in South America (Kingstone, 2011) . Although the percentage of the GDP accounted for by the industrial sector has declined over the last decade, industrial manufacturing stil l accounted for an impressive US $231 billion in 2012, more than the manufacturing
22 sectors of Argentina, Chile, Colombia and Venezuela combined. 6 Furthermore, the sector averaged 2% annual growth between 2004 and 2012. Like most countries of the Global So uth, Brazil industrialized later than the US and Europe. Historians frequently attribute this to the 19 th century export oriented oligarchy, which dominated the political economy until the late 1800s. This group, which mostly consisted of the landed elit e, maintained its privileges and position through the control of resources and the prevailing agricultural economic system built on the foundation of slave labor (Bueno, 2005) . Such individuals proved resistant to early attempt s to modernize the economy. 7 slavery and the adoption of a Republican form of government. Although manufacturing led industrializatio as an outgrowth of the SÃ£o Paulo coffee economy. Logistical concerns of coffee planters and exporters led to the creation of an extensive railway system in SÃ£o Paulo state, and demand for agricultural labor attracted signifi cant immigration to the southeastern region. Immigrants that were originally recruited to work on coffee plantations frequently opted to find work instead in the increasingly urban capital, leading to a labor surplus in the area. An abundance of cheap labo r in SÃ£o Paulo made 6 . According to data provided by CEPAL, the Brazilian manufacturing sector decreased from 16% in 2004 to only 11% of GDP in 2012. 7 . Souza, sought to industrialize Brazil and supplant the economic system of the time with a model resembling the industrialized na tions of Europe. He founded and ran more than a dozen businesses during the reign of Dom Pedro II, and incessantly promoted liberalization of the economy. Tragically the landed elite, and the emperor proved recalcitrant to his efforts, and his opponents su cceeded at bankrupting him through financial stratagems and sabotage. His tragic saga is a testament to the (Bueno, 2005) .
23 modest, mostly limited to textiles, footwear and foodstuffs. With the onse t of the great depression and decline in demand for agricultural exports, the Brazilian government adopted macroeconomic policies as emergency refer to this time as an e conomic revolution in Brazil, where we see the rupture of the cafÃ© com leite (Sader, 2010 12, 15). Also during this period, Brazil b egan its rapid expansion of the industrial sector with marked success. The federal government implemented protectionist policies such as increased tariffs on manufactured items, in order to encourage domestic manufacturers. Such policies correspond to the ISI (import substitution industrialization) model that characterized Latin America in the decades following the Great Depression and World War II. contributed to rural to ur ban migration as migrants left the countryside to move to industrial centers in search of work (Kingstone, 2011) . This allowed large portions of the population who previously lacked basic services to gain access to education an d healthcare. 8 Also during the 1940s and 50s Brazil witnessed rapid growth of its middle and industrial working classes, two groups that came to play significant roles in the politics of the early 1960s. Perhaps the most impressive legacy of ISI however i s the rapid economic growth Brazil experienced during the period. The Brazilian GDP grew at 8 . According to historical data provided by BÃ©rtola & Ocam po, the Human Development Index in Brazil improved significantly between 1950 and 1980 increasing an average of 30% per decade.
24 an average rate of 7% between 1948 and 1980, peaking at 13% in 1973 during the so Morandi, 2003: 7). Despite these su ccesses however scholars have been quick to point out problems associated with ISI, such as the growth of the informal sector, national debt accumulation, bloated bureaucracies and inefficient state run enterprises (SOEs). With the global fuel crisis in 1 973 and skyrocketing interest rates, Brazil, like other nations in Latin America, rapidly accumulated debt. According to economic historian Bulmer Thomas (2003), Brazil was among the nations of the region to look to means of maintaining growth during a time when world demand for primary materials was in decline. Nonetheless by the 1980s, national debts had reached such levels that most Latin American countries could no longer realistically service their debt payments . After Mexico declared it would default on its debt servicing payments in 1982, a general lending crisis ensued with lenders restricting loans to the region. Subsequently, strategists from the IMF provided a plan for the restructuring of national debts, w hich required countries to follow a series of policies, known as structural adjustment (SAPs) (Bulmer Thomas, 2003) . Although the Brazilian government loosened many of its protectionist policies during the 1990s as a means to c urb inflation and reduce the national debt, to this day it remains a relatively protected economy. Perhaps for this reason, the manufacturing sector in Brazil has continued to contribute a significant percentage of GDP, 11.3% (Alice , 2013) . Despite relatively high labor costs, Brazil is nearly self sufficient in many manufactured items, including foodstuffs, petroleum products, clothing electronics and automobiles.
25 Gender and Development The 1970s witnessed a flourishing of fem inism and gender theory from scholars and researchers, which heavily influenced the way policy makers think about development today. Gender equality is now at the forefront of many discussions on development and has been a major focus of the United Nations and other institutions committed to human progress and social justice. 9 Many policy makers have recognized that poverty and other development issues disproportionately affect women and children, thus have rationalized that by implementing programs that ta rget women, they might inadvertently reduce infant mortality, malnutrition and other development indicators. Such an approach has been criticized, however, by feminist scholars who have argued that the empowerment of women as individuals is a worthy goal i n and of itself, and should not necessarily be conflated with other development goals (Chant, 2012) . Regardless of the motives which drive institutions and individuals to make that empowering women is good for development (Duflo, 2011) . discourse and academic journals, its definition remains somewhat vague. Such is the position taken by Kabeer (1999) who suggests that the value of the term for many power, over their own lives (Duflo, 2011; Kabeer, 1999) . Aiding women to gain their own 9 . population, and are generally exacerbated by gross inequality between men and women.
26 s concerned with development, if for no other reason because it makes good economic sense. The literature demonstrates that women tend to spend more of their income on their children than do men, and that the educational attainment of mothers is more posit ively (World Bank, 2012) . A variety of indices and statistical tools were created by the United Nations Development Programme (UNDP) to assist in the study and assessment of women's empowerment and gender equality. The Gender Inequality Index (GII) is among the most comprehensive of these indices, since it is a composite of various indicators related to reproductive health, empowerment, and labor market participation. Like the HDI, the GII is useful tool for comparing the level of development among different countries. In 2012, Brazil ranked number 85 with a GII value of 0.447, placing it within the bottom half of 148 countries indicating there is much r oom for improvement. 10 The Feminization of Labor Esther Boserup (1970) was among the first economists to considered gender in Development, she showed the diversity of productive ta sks in which women regularly a western male centered paradigm are often unsuccessful because they fail to take into 10 . Gender Inequality Index, da ta available United Nations Development Programme: Human Development Reports. < http://hdr.undp.org/en/data>.
27 ricultural and reproductive which would become a significant issue only a few years after the publication of her book, many of her observations and conclusions remain relevant. One of the characteristics of the 1970s was a rapid feminization of the industrial workforce, which was particularly pronounced in less developed countries undergoing export oriented industrialization. An abundance of scholarship subsequently e merged addressing the gendered composition of the workforce in export processing zones, and industrial centers in the Global South (Hein, 1986; BenerÃa L. a., 1987; Pearson, 1986) . These scholars often ref erence the so notion among those employed in manufacturing that women are physically better suited to manually intensive assembly tasks, because of their anatomical features (Pearson, 1986). Many researchers are skeptical that employer preference for female employees is actually rooted in physical differences but rather because, in practice, female workers tend to be more submissive, easier to dismiss, and willing to accept lower wages, due to gender socializati on and roles (Hein,1987). The modern notion of equal work for equal pay has often been ignored by factory workers at an electronics manufacturing plant in Mauritius she found that employer preference for females was overwhelming because of Mauritian legislation mandating higher pay for male workers. While this type of legislation, was originally meant to insure a living wage for men and their families (since men at the time wer e generally viewed as household heads) it advertently contributed to male unemployment.
28 Hein noted that on occasion there were labor shortages of female workers alongside high male unemployment rates. Injustices perpetrated by employers against women work ers are not limited to wage discrimination. The literature on gender discrimination in the workforce is replete against them becoming pregnant, or forcing birth control on their workers, are all too frequent (Roberts, 2003; Torres, 2005) . In other circumstances pregnancy is welcomed by employers as it fac ilitates turnover among workers helping to keep wages down. A study by GoldÃn (2011) of labor turnover rates in export processing zones in the highlands of Guatemala, found that employers often preferred young female workers because they were more likely t o terminated their own employment due to pregnancy, whereas male workers, particularly older ones, were more likely to remain with the company for an extended period, putting upward pressure on wages. The Defeminization of Labor In the early 1980 s feminist researchers began discussing a new and initially surprising trend with regards to female employment in less developed countries. After approximately two decades of unprecedented growth in the female labor force spurred on by rapid export oriented industri alization, progress made towards the inclusion of females in industry began to reverse itself (Chant, 2003; Pearson, 1986) . Industries, such as electronics manufacturing and metallurgy experienced a reduction in the proportion of females they employed (Pea rson 1990). Scholars have characterized this
29 implication for gender equality and human development. Inherent in restructuring is the concept of flexible accumulation, whereby businesses implement cost saving mechanisms to maximize profits, which then can be reinvested to maintain competitiveness. In the economies of Latin America with large manufacturin g sectors, this includes the adoption of more capital intensive forms of manufacturing and the reduction of labor intensive tasks. Feminist economist, Ruth Pearson noticed this phenomenon when studying maquilas along the US Mexico border, during the 1980s. At that time, many Mexican factories in the garment and electronics subsectors began modernizing their production processes to remain competitiv e in the changing economy. The defeminization of labor, is tied to such modernization, as capital intensive pro duction methods come to replace manual assembly processes. As businesses update their technology, a single new machine may replace scores of assembly workers. Thus tasks that were previously carried out by multiple unskilled workers come to be performed by a single skilled machine operator resulting in layoffs as firms eliminate non critical positions (Pearson, 1990). During this period women's employment as a proportion of total employment in the maquilas along the border region fell by 3%. Pearson observ ed a more pronounced decrease of as much as 20% in subsectors that already had a large male presence, i.e. those that produced metal products, furniture, and transportation (1990:155). Pearson identifies the trend towards the masculinization of certain sub sectors as yet another
30 Conclusion Brazil has followed a course of development similar to other nations of Latin America, although it industrialized relatively early for the region, and maintains one of the mos t protected economies in the hemisphere. Today, diverse products are manufactured in Brazil for domestic and foreign consumption. The bulk of such manufacturing takes place in a few high population centers in the southeastern region of the country, in the states of Minas Gerais, SÃ£o Paulo, and Rio de Janeiro. These few states together have historically accounted for most of Brazilian GDP, although in recent years there has been a slight decline in their participation. 11 More recently, the federal and state governments, aware of the benefits that industrial jobs bring to the southeastern population, have encouraged the decentralization of the industrial sector by extending incentives for businesses to relocate to less developed regions of the country. This ha s given rise to a number of industrial districts throughout North and northeastern Brazil, including Salvador, Recife and Manaus (Gaspar, 2003; Veja, 2012) . The Brazilian government has devoted substantial resources t o human development issues over the last several decades and made significant progress in improving certain indicators, including literacy, reduction of infant mortality, and poverty alleviation. As the largest country in Latin America, it provides an inte resting context through which to study the relationship of women's empowerment and economic development. Brazil has often been at the forefront among Latin American countries in 11 . According to an article published in Veja Magazine in 2012, the Southeast region accounted for 55% of the Brazil GDP in 2010, a decline of 1.5% from 2002 (Veja, 2012).
31 guaranteeing human rights, including those related to gender. 12 There are numerous federal and state institutions that promote gender equality, and multiple laws that protect women from wage discrimination and intimate partner violence. 13 In C hapters 2 and 3 we address more specifically the trajectory of female workers in the industries of Manaus. 12 . Brazil has been at the forefront of extending secondary human rights, some examples are the Cardoso administrations highly successful campaign against AIDS and its promotion of indigenous rights and other land rights of tradi tional peoples. 13 . In of Women's Rights are federal institutions, similar organizations are also present at the state and municipal levels. Among the laws designed to protect women against discrimination is article Constitutional Law 11.340, 2006. Also known as lei Maria da Penha is designed to p rotect women against intimate partner violence by stipulating severe punishment for perpetrators of violence against women. Additionally, the federal constitution prohibits wage discrimination based on gender.
32 CHAPTER 2 THE MANAUS FREE TRADE ZONE: ORIGINS AND CHALLENGES Manaus is an excellent case study of state fueled development because its entire economy is dependent on a man ufacturing sector, conceived and maintained through tax incentives. It thus provides a relatively concise and well documented example of state led industrialization, which has garnered considerable attention from both Brazilian and foreign scholars. Inter five years in light of macro economic and political changes causing businesses in the Manaus Industrial District to restructure their management and production models (Andrade et a l . 2012: 181). It is in the context of these changes that I examine employment in the Manaus Industrial District. Here I review the literature on the Manaus Free Trade Zone and the phenomenon of restructuring, the sexual division of industrial labor and t he role of unions. It is divided into two main sections: In the first section I review the history of the creation, as well as studies on factory workers. In the following section I review the literature on restructuring with regards to its effect on the workforce and labor market of Manaus. I conclude by summarizing the main findings and emphasizing current lacunas in the literature. of the Manaus Free Trade Zone as an example of state led development. Particularly
33 important, are its history of economic cycles, and understanding government policies in the Amazon prior to the arrival of the free trade zone. The Amazon: A History of Economic Cycles and Fiscal Incentives Many critics of the MFZ view it as a departure from other development projects because of its total dependence on fiscal (or tax) incentive s (Pereira 2005). Furthermore they question the model's effectiveness because of the artificial, enclave like economy it induced (Pereira, 2005; GonÃ§alves, 2012). Unfortunately, such views often fail to take into consideration the historic role of fiscal i ncentives in spurring other economic cycles in the Brazilian Amazon. Historians of the Brazilian Amazon usually refer to the "spice cycle" as the first major economic cycle in the region. Beginning in the 17th century, Jesuit missionaries established missi on settlements ( aldeias ), where they evangelized the natives and oversaw communal agriculture compounds. These catholic fathers supported their settlements largely through the sale of cultivated foodstuffs and forest products, the latter being d rogas do s ertÃ£o (or spices). The Portuguese Crown provided them with substantial fiscal incentives at the time to encourage the cultivation and export of spices, which became a lucrative industry (Cleary, 2004; Ravena, 2005). After the expulsion of the Jesuits in th e mid 18th century, fiscal incentives for agriculture and the extraction of spices continued under the reign of the Portuguese viceroy, JosÃ© SebastiÃ£o Carvalho de Melo (Marques de Pombal). Pombal founded The General Company of Commerce of GrÃ£o ParÃ¡ and Mar anhÃ£o which the Portuguese crown exempted from export taxes and gave exclusive right to the export of all goods originating in the captaincies of GrÃ£o ParÃ¡ and MaranhÃ£o (Ravena, 2005: 134). Although the population of the Brazilian Amazon grew during this p eriod with the influx
34 of farmers, spice collectors, and African slaves, the fiscal incentives enjoyed by Pombal's company did little to increase human development in the region and most inhabitants lived in penury (Cleary, 2004: 117). 1 Even centuries later region continues to lag behind the South, Southeast and Central west regions in terms of human development and infrastructure (Lemos, 2007) . Although it brought little and led to the formation of riberinho culture. 2 Even more influential in the formation of the region than the "spice cycle," however, was the "Rubber Boom" of the late 19th century. By 1839 Charles Goodyear i nvented the process of vulcanization, leading to the demand for rubber as an industrial product (Ferreira, 2003: 17; Souza 2010: 3). Because the Brazilian Amazon contained the majority of the world's h evea brasiliensis (rubber trees) at that time, it witne ssed one By the 1870 2010). This brought an unprecedented amount of economic growth and prosperity to the region, as well as a dr amatic increase in population (Cardoso, 1977) . Although most of relatively small group of rubber barons and exporters, there was some investm ent in public works during the period. It was during this time that Manaus received its famous 1 . It is during this period that we see the fo rmation of the Cabano class, an ethnic and social group discussed at length by Cleary (2004) . As their name implies Cabanos were a socially marginalized group that lived in rudimentary housing or "cabins." Their very formation exemplifies the high degree of economic and social inequality that characterized the "Spice Cycle . 2 . Riberinho refers to a diverse group of Amazonian fluvial dwellers, who mostly subsist on fishing and small scale agriculture (Freitas, 1999)
35 opera house ( o Teatro Amazonas ) and became one of the first Brazilian cities to have electric lighting and a trolley system (Hemming, 2008; Souza, 2010). Despit e such investments however, the vast majority of Amazonians continued to live in poverty, many of them in a state of indentured servitude imposed by the aviamento labor system. 3 Although per capita GDP increased during the period this did little to minimiz e inequality or encourage human development. 4 The decline of the "Rubber Boom" is well documented, as are its effects on the population and economy of the Western Amazon (Cardoso, 1977; Hemming, 2008). Due to new competition from rubber grown in Southeast Asia, Brazil went from supplying more than 60% of world demand in 1910 to only 2% by 1919 (Cardoso, 1977: 28; Mahar, 1976: 148). The Brazilian government adopted various policies aimed at saving the rubber extraction industry, and preserving the way of l ife of thousands of rural Amazonians, during this time . As early as 1912 the federal government began promoting through subsidies the creation of rubber stands, or plantations of hevea brasilensis , in an effort to increase the supply of latex and reduce the labor associated with its exaction . Also during this period the government eliminated tariffs on the importation of all tools associated with the extractive industry and instituted a system of subsidies for natural latex (D'Almeida 1982, Mahar 1976). The most aggressive attempt to resurrect the industry, however, 3 . Under the aviamento system, the rubber tapper could only sell the latex he produced to a single landholder or rubber baron. Further more he was required to purchase all items for his sustenance from a single trading post, controlled by the landholder. Because of exorbitantly high priced goods at the trading post, and prohibitions against practicing any agriculture, the rubber tapper us ually accumulated great debt. Which prevented him from making any profit and achieving social mobility. 4 . Although there are no estimates available of the GINI coefficient for the population of Manaus during the historical period in question, based on the available literature one can assume that there was pronounced inequality just as in other urban centers during the era (Hemming, 2008)
36 was not until the 1940s with the Acordos de Washington , an international agreement between the Brazilian and United States federal governments. This World War II agreement designated Brazil as the sole supplier of rubber to the US from 1942 to 1947 and although it generated substantial activity, peak production only achieved a fraction of what it was during the rubber boom (Pereira, 2005: 9 10). Fortuna tely, by the conclusion of the Acordos de Washington the demand for rubber in Brazil's own industrial sector outstripped international demand and thanks to protectionist tariffs on foreign produced latex the rubber extraction industry survived (Schmink and Wood, 1992: 48). Such protectionist polic ies were hard won, however, by those with an invested interest in the economy of the Western Amazon. In 1953, federal deputy Francisco Pereira da Silva delivered an impassioned speech before the Brazilian congress imploring that the government reconsider t he importation of synthetic latex because of its detrimental impact on the national rubber extraction industry. He implored the president and members of congress to consider the t 2005: 101). SerÃ¡fico (2005) suggests that this move by da Silva, at a time when Brazilian industry was rapidly expanding under the ISI development model, exemplifies the sentiment of many local pol iticians who hoped to integrate their region to the rest of the Brazilian economy during a period of rapid growth. Creation of the MFZ Da Silva was the most important proponent in the creation of the Manaus Free Trade Zone. Although a native of the Northea st, he spent most his political career in the North, and continually championed the development of the Amazon (Pereira 2009). He
37 was the first to propose the Manaus Free Trade Zone as a model for regional development, authoring multiple pieces of legislati on over 17 years that culminated in the Manaus Free Trade Zone in 1957, and its reorganization as an industrial model in 1967. 5 Federal law 288 (Feb 28, 1967) extended the geographic area of the MFZ beyond its original 200 hectares to 10,000 km 2 and estab lished robust fiscal incentives for the companies operating therein (Botelho, 2006; SerÃ¡fico, 2005). The law also called for the creation of Suframa, an autonomous governing agency that works in collaboration with the state and municipal governments (Filho , 2006). 6 The Manaus Free Trade Zone was meant to spur economic growth in and manufacturers began operations in Manaus, this would lead to an array of commercial ac tivities, with factories establishing forward and backward linkages with other businesses in the region (Mahar, 1976). Revenues generated through the model could be reinvested in the region, and the population would benefit from increased employment opport unities and a rising standard of living. exemption from taxes on industrial products (IPI) and exemption from taxes for export of products manufactured in the Zone ( Suframa/Histori a ). Companies located within its limits, receive the best tax breaks in the country. Considering that Brazil is South 5 . In 1951 Da Silva presented law project 1.310 which proposed the creation of the free port of Manaus. This project w as eventually modified to become law 3.173 in 1957. It was further regulated by decree 47.754 in 1960 and finally came into effect in 1967 with law 288 28.2.67 (SerÃ¡fico, 2005: 101). 6 . Suframa: An acronym derived from the Portuguese, SuperintendÃªncia da Zona Franca de Manaus (Superintendence of the Manaus Free Trade Zone).
38 America's most industrialized nation with one of the most burdensome tax codes in the hemisphere, it is no surprise that this nearly tax f ree enclave is a source of considerable debate (Alper, 2012; The Economist, 2013). Opponents of the MFZ claim that such an unequal distribution of the federal tax burden is unjustifiable, because the model has failed to bring the degree of development that it was originally intended to (Pereira, 2005; Bomfim, 2009). 7 Although most of its critics historically hale from the industrial Southeast, an increasing number are found locally within academia (Maciel, 2009; Ferreira, 2003). The Free Trade Zone and Re gional Development Among the most prevalent criticisms of MFZ as a development model is that it does not benefit populations living outside of Manaus. As early as 1976, Mahar nt and development outside the immediate vicinity of the city (1976: 170). By examining regional statistics relevant to education, employment and income, Pereira (2005) argues that human development in the state of Amazonas significantly lags behind that o f the state capital. She provides strong evidence that the MFZ does not directly benefit the rest of the Western Amazon in terms of economic and human development and thus has failed to fulfill its original mission of bringing development to the entire reg ion. These arguments are especially relevant when we consider that according to its charter, the Manaus Free Trade Zone is meant to be a regional and not just a municipal development project. 7 . According to the official discourse the mission of the Manaus Free Trade Zone is to promote "...economic development ... integration of [the Western Amazon]... [and] provide a better quality of life for the populations therein" ( Suframa/Projeto )
39 Wesche and Bruneau (1990) suggest that during the 1970s and 198 0s the MFZ Itacoatiara. They found that, during the first two decades of the project, the allure of and educated individuals promptly moved to Manaus in search of employment in the manufacturing sector. Their observations with regards to Itacoatiara likely apply to other municipalities in Amazonas, since the rate of rural to urban migration throughout th e state dramatically increased after creation of the free trade zone. 8 Similarly, Mahar (1976) suggests that the Manaus Industrial District acts as a "pole of attraction" rather than a "pole of radiation" attracting foreign businesses and migrants to the c ity, without stimulating economic development elsewhere (154). Much like the rapid urbanization experienced by other Latin American nations as they underwent ISI, city planning and investment in public works failed to keep up with pulation. Luiz Botelho (2006) points out that the expansion of employment opportunities, housing, and adequate sanitation in Manaus have not kept pace with the demographic explosion of the city. Botelho's comments are consistent with those of DesprÃ©s (1991 ) who found that despite the positive publicity surrounding characteristics of a "huge favela or slum" (3). 8 . According to official IBGE statistics the population of Manaus in 1970 was approximately 313,000, then increased in 1996 to 1,500,000, increasing further to 1,6 44,690 by 2000 (Barbosa 2007: 57). According to the 2010 census the population of Manaus reach 1,802,014, more than half of the inhabitants of the state of Amazonas. This represents an increase of the population of Manaus by approximately 28% while the tot al state population increased by less than 20%.
40 Macro Economic Policies and the MFZ The Manaus Free Trade Zone as a development model can best be understood in the context of Import Substitution Industrialization (ISI) whereby governments enact protectionist policies in order to encourage domestic industrialization (Kingstone, 2011). In Brazil such policies have h istorically included high import tariffs on foreign manufactured products in order to assure the competitiveness of domestically produced goods (Kingstone, 2011: 29). Although protection allowed Brazil to become self sufficient in many manufactured items, including automobiles and electronics, such products were often more expensive and of lower quality than similar goods produced elsewhere. This is because products manufactured in ISI industries often include inflated production costs, which are passed on to consumers (Smith, 2009). Taxes also add to the final cost of domestically produced items in Brazil. For certain goods such as electronics, automobiles, fine wines, and tobacco considered luxuries by the Brazilian tax code cumulative taxes can incr ease the final price by the Manaus Free Trade Zone have a significant advantage as the fiscal incentives they receive dramatically reduce their tax burden giving them a comp etitive edge over businesses located elsewhere. Unlike export processing zones or free trade areas, the MFZ does not specialize in assembling products for export. 9 Rather, 95% of all items produced by factories located at the MFZ are for domestic consumpt ion. Thus as long as protective tariffs on 9 . The ColÃ³n Free Zone ( Zona Libre ColÃ³n ) in PanamÃ¡, with its more than 1,600 companies is perhaps better known than the MFZ, however it is very different model. For one thing the ColÃ³n Free Zone, specialize s in commerce rather than industry, that is, goods are imported duty free and resold to foreign consumers and firms. The Manaus Free Trade Zone on the other hand specializes in manufacturing of value added goods the majority of which are sold to Brazilian consumers.
41 imported goods and taxes on domestically manufactured industrial products remain elevated elsewhere in the country, manufacturers in Manaus remain competitive. Any reduction in protective tariffs or taxes elsewher e in the nation, however, can have an adverse effect on the competitiveness of firms in Manaus, to the detriment of the local workforce. Neoliberal reforms in the early 1990s negatively affected the MFZ model. Although such reforms arrived later in Brazil and were less robust than in other Latin American countries, their adoption signified a departure from the ISI model of development. The period was marked by decreases in government spending, the privatization of SOEs (state operated enterprises), and the gradual opening of the Brazilian economy (Costa, 2005; Kingstone, 2011; Roett, 2011). At this time foreign manufactured goods became more affordable for Brazilian consumers, as did products manufactured in the industrial Southeast. Subsequently, productio n in the Manaus Free Trade Zone suffered as many factories were forced to downsize ( Suframa/Historia ). According to official statistics published by Suframa, between the years of 1990 1992 annual revenues in the Manaus Industrial Pole decreased from USD$ 8.3 billion to just USD$ 4.5 billion (Suframa, 2010: 14). This signaled a reduction in production and employment as tens of thousands of workers lost their jobs. The MFZ as an Employer The right to work is enshrined in the 1988 Brazilian constitution. In Article VI, work is listed as a basic social right, alongside rights to housing, food, education and healthcare. The inclusion of the right work in Brazil's supreme legal document provides for an interesting debate surrounding the importance of work in Bra zilian society and the role of the government in curtailing unemployment.
42 When we consider the constitutional obligation of the Federal Government to generate employment opportunities in Brazil, the Manaus Free Trade Zone takes on a special significance be cause its industrial district, created through state intervention, remains the main employer in the city (Scherer, 2004: 135). Although employment numbers at the Manaus Industrial District have fluctuated over the last four and a half decades, today the di strict directly employs more than 100,000 individuals and provides another estimated 400,000 jobs indirectly ( Suframa/Desenvolvimento_Regional ). There is no guarantee, however, that such levels will continue. Should domestic demand for goods produced in th e Manaus Industrial District decline, it will likely lead to a decrease in employment. 10 Although initially the Manaus Free Trade Zone was able to absorb the steady onslaught of jobseekers migrating to the city, this no longer appears to be the case. In h is 1976 study of the MFZ, Mahar points out that between 1968 and 1971 unemployment in Manaus fell from 9 to 8.2%, a substantial decrease, considering the sector quickly expan ded alongside its manufacturing sector, providing an abundance of jobs in commerce and services in addition to those available in industry. By the 1980s Manaus had become a popular tourist location for Brazilians, with thousands of visitors flocking there annually in pursuit of low cost duty free electronics (Mahar, 1976; Suframa/Historia manufacturing and commercial sectors suffered and its tourist industry dissipated. 10 . In spite of the Free Trade Zone, unemployment and sub employment remain pressing issues in Manaus. In 2010 for example the unemployment rate in Manaus was 11.1 %, making it one of the highest among the 20 most populous cities in Brazil. Tha boasted a record number of employees, suggesting that labor market supply far exceeds demand even at peak productivity (Cruz, 2011).
43 Although the popul ation of Manaus has continued to grow through migration and its manufacturing sector has expanded in recent years, the relative demand for labor is Assembly Workers Although one can partially assess employer through employment figures from Suframa, such data reveal little about job quality. For this reason, scholars of the MFZ, interested in issues of human development, equality, and empowerment, often empl oy qualitative methods into their research, such as interviews with workers (Barbosa, 2007; Carvalho, 2011; DesprÃ©s, 1991; Scherer, 2004). US based anthropologist Leo DesprÃ©s (1991) conducted an early study using qualitative methods to examine the lives an d working conditions of unskilled labor in the Manaus Industrial Pole. To fill a lacuna in the literature with regards to urban employment in Manaus, DesprÃ©s interviewed 205 individuals from various sectors including the manufacturing, public, commercial and informal sectors. 11 DesprÃ©s's sample was large enough for him to create a general profile for electronics assembly workers and assess how their job quality and levels of satisfaction compared to those of individuals employed in other sectors. He found t hat the majority of assembly workers were female, young (between the ages of 18 and 26), high school graduates, and disliked their job. 12 The assembly workers he interviewed generally expressed less satisfaction with their jobs than did 11 . DesprÃ©s distinguished between, four different employment sectors in his study Work and Social Life in Brazil's Free Trade Zone: Hegemonic Industrial, peripheral industrial, commercial and public sectors, and the informal sector. His findings are especially revealing because he was able to compare workers' attitudes, incomes, profess ional qualifications, and household dynamics between different employment sectors. 12 . He only interviewed assembly workers from the electronics subsector in two different firms.
44 individuals employed in the commercial and public sectors. The main reasons they gave for their dissatisfaction were low wages, long hours, the tedious nature of assembly work, and the lack of opportunity for social interaction with coworkers and supervisors (DesprÃ©s, 1991: 1 12 113). DesprÃ©s found that although assembly workers averaged higher wages than employed in the service and public sectors. He suggested this is one of the main reasons assemb ly workers often consider their work in the MID as transitional employment rather than as a long term career path. Similar studies have pointed out locations in the state a nd country employing workers of similar skill levels (Nunes de Miranda, 2013) the earnings of the equivalent subsectors in SÃ£o Paulo is especially revealin g. In 2012 the mean monthly earnings of an electronics manufacturing worker in the MID were BR$ 1539.76 compared to BR$ 3107.00 in SÃ£o Paulo ( Table 2 1 ). While we should be cautious about directly comparing these figures since they come from different sources, and there are other factors to consider such as cost of living differences between the two locations, they suggest that businesses in the MID spend significantly less on labor than their counterparts in the Southeast. As M iranda (2013) suggests, such low wages in the Manaus Free Trade Zone are curious when we consider the already unparalleled savings businesses receive through tax incentives (13). Although the literature is unclear on why firms in the MID have been able to keep wages relatively low it may be due to a seemingly inexhaustible supply of unskilled and semiskilled labor in the city.
45 Female Workers While the Manaus Industrial District has been a major employer of women since its inception, Suframa only began prov iding employment figures disaggregated by gender in 2004. Accordingly, the female share of the workforce peaked at approximately 35% in 2005 fluctuating thereafter between 34% and 29% ( Table 3 1 ). In 2013, 31.6% of the workforc e was female indicating a general decline of more than 3% since 2005. One possible reason for this decline in the share of jobs occupied by women in the MID may be due to the rapid expansion of the Bolsa FamÃlia program over the last decade. Researchers in dicate that Bolsa FamÃlia recipients in urban areas are likely to remain unemployed longer than those that receive no assistance, decreasing the labor supply (Ribas, 2011) . In 2005 less than 2% of eligible families in Manaus received assistance through Bolsa FamÃlia ; however, this figure steadily increased, reaching 59% of eligible families in 2013 (MS/SAS/DAB/CGAN, 2014) . Given that the majorit y of recipients are women with children, it is not unlikely that the increased coverage of the program over the last decade has reduced the female share of the labor supply, hence related to the decline in the share of jobs occupied by women in the MID (Secretaria de PolÃticas para as Mulheres, 2013) .. Although Suframa does not disaggregate their workforce data by subsector, one can assume that the gendered distribution of the workforce is not uniform across industries (Suframa, 2013: 98). This is because the literature on the MFZ indicates that women are overrepresented among the workers in electronics manufacturing and a rarity among workers in the vehicle subsector (Barbosa, 2007; DesprÃ©s, 1991; Social, 2001; Spindel, 1987).
46 Various scholars have studied gender in the context of the electronics subsector in the Manaus Industrial District (Lima, 2009; Spindel, 1987; Torres, 2005). Among these is IerecÃª Barbosa (2007) who surveyed 895 assembly workers from six different electron ics manufacturing plants in Manaus in 2001 (47 48). She used both quantitative and qualitative methods, including interviews where she questioned participants on domestic vi olence and sexist rhetoric, reproductive rights, and computer literacy. 13 Along with these questions she collected extensive demographic data for each participant, including age, marital status and history, motherhood, education, number of previous jobs, re ligion etc. This allowed her to assemble the most comprehensive profile to date of female electronic assembly workers in Manaus. Among her most interesting findings are that approximately 8.5% of those interviewed had at one time served in a leadership po sition while employed at the Manaus Industrial District, and the majority of those interviewed had children and were partnered. 14 Her findings regarding motherhood were particularly interesting because the literature on the feminization of factory labor els ewhere indicates that employers generally prefer females without children (GoldÃn, 2011). The reason such a large percentage of the female workforce in Manaus has children is unclear. This may simply be due to demographic characteristics of the young femal e adult population in Manaus, 13 . ir own lives. Thus potential, and secure autonomy from individuals and situations that limit their individual freedom and choices, is essential. Computer literacy re provides individuals to information and employment. Nearly half of those Barbosa interviewed indicated that they were comfortable navigating the internet (99). 14 . iage like relationship.
47 or may indicate the success of labor laws prohibiting gender discrimination and childcare funded by the employers in the district. Barbosa concludes that the management structures of the firms in her study perpetuate patriarc She found that despite some economic autonomy that wage employment provides these women, they remain limited by their family structures and lack of opportunities for advancement within the companies they work for (136 137). She suggests that a liberal education rather than vocational training is the best method to empower women as it they will learn to question and challenge the prevailing patriarchal structure that marginalizes them (137 1 38). Restructuring Central to our understanding of changes in the share of female workers and labor flexibility in the Manaus Industrial District, over the last 40 years, is the concept of restructuring. Although the precise definition of restructuring va ries depending on the context, in our discussion of employment at the MFZ it will refer to a generalized process whereby firms adapt to macro changes in the political and or global economy. adoption of more automated processes and increased use of flexible forms of labor. These changes are implemented by firms with the goal allowing for quicker response to fluctuations in market demand an d decreasing production costs. The phenomenon of restr ucturing has had a major impact on the workforce of the Manaus. Fordism as the Traditional Industrial Model In order to better understand changes in the industrial workforce brought about by restructuring, it is worth reviewing the prior industrial model that dominated
48 production. Fredrick W. Taylor and Henry Ford are credited with the creation and implementation of the production model that characterized manufacturing in industrialized nations for most of the 20th century. As an engineer, Taylor pioneered the scientific study of the division of work tasks. He discovered that producers could greatly enhance efficiency by dividing the manufacturing processes into clear, repetitive tasks and by keeping production apart from planning and administrative process es. As an production of automobiles. Today the terms Taylorism and Fordism are often used interchangeably to refer to 20 th century production models that include a se gmented hierarchical structure, in which workers carry out clearly defined roles with specific tasks and responsibilities (Souza, 1998; Vidal, 2008) . Much of the literature depicts Taylorism and Fordism as conducive to amiable relations between employers a nd employees. Under such conditions the excessive relations, to 2008: 921). Additionally, Fordism/Taylorism is seen as compatible with union recognition and collective bargaining, structures that benefit the workforce (Souza, 1998: 6). For these reasons scholars concerned with workers issues and declining job quality brought about by restructuring, tend to make nostalgic references to Fordism (Ramos, 1997). Flexible accumulation and Toyotism Brazilian scholar of labor law, Alexandre Ramos (1997), identifies 1973 as the year global economic conditions shifted away from favoring the Taylorist/Fordist schools of management. He suggests that the world petroleum crisis and accompanying
49 produced goods, leading to the unsustainability of traditional models of mass production (76) . He suggests that during this period US and European automotive manufacturers began to gradually incorporate management techniques that were developed by Japanese manufacturers during the 1950s and refer to management styles modeled after these Japanese techniques, the latter favoring batch production of varied goods over the less flexible mass production of a few items. 15 less segmented division of work tasks, and high usage of subcontracting and outsourcing of secondary and tertiary processes (83). Toyotism tends to value multifaceted assembly workers since processes that were handled by a separate departme nt under the Fordist model, become the responsibility of workers on the production line. According to Ramos (1997), T oyotism is a highly flexible model in terms of labor, ributes to the accumulation of capital, which generally benefits only a few stockholders in the form of profits. He is critical of Toyotism in how it affects workers, suggesting the model is characterized by lower wages with less benefits, weaker unions, a nd higher turnover rates. He also claims that the model results in an increased polarization of workers between a limited number of high status skilled positions, and a large number of low status unskilled positions. 15 . Economist Harold Perkin (1996) referred to this global shift away from Taylorism as the third revolution (as subsequent to both agricultural and industrial revolutions). Other scholars have referred to this new model as the thir d wave, post fordism, the automation revolution, and flexible accumulation. Many Brazilian scholars actually prefer the term "Toyotism," since it is analogous to "Fordism" thus allows for convenient comparison (Ramos 1997).
50 Opening of the Brazilian Economy Altho ugh Brazil was affected by the fuel crises and other macro economic trends manufacturing sector than elsewhere in Latin America. This is largely due to the onist policies as a holdover from its inward looking phase of development. Up until the early 1990s the Brazilian government maintained relatively high tariffs on imported goods and had an abundance of state operated enterprises (SOEs). In an effort to cur b inflation the Collor administration began liberalizing the Brazilian economy, by opening it up to foreign investment and privatizing SOEs (Roett, 2011). Although at the time, neoliberal reforms were championed as a means to reduce inflation and build th e economy, many scholars have questioned how they affected human development (Costa, 2005; Kingstone, 2012; Ponchmann, 2009). BÃ©rtola and Ocampo (2012) acknowledge that while inflation radically declined in Brazil following the implementation of reforms in the 1990s, unemployment rates increased during the same period. Perhaps nowhere was the loss of jobs more pronounced, however, or better documented than in the case of the Manaus Free Trade Zone. Restructuring and Precarious Employment in Manaus In her study of work in Manaus, anthropologist Elenise Scherer (2004) discusses the effects of restructuring on employment. She concludes that job quality and employment stability generally declined after the opening of the national economy in 1990 and subsequent restructuring. She interviewed scores of unemployed and sub employed individuals between 1999 and 2000, many of whom had previously worked in the manufacturing sector for businesses in the Manaus Industrial District. She found
51 that many of these individua ls had lost their jobs due to routine lay offs and despite sustained efforts were unable to secure new employment at the district. Many of her research participants reported being forced to take jobs in the informal sector where they generally received les s income without the benefits prescribed by Brazilian labor legislation and local labor unions. 16 Scherer equates the growth of informality to a are unable to contribute to so cial security and thus lose their rights to retirement 2004: 139). It is worth noting that although there is currently no legislation in Brazil requiring employers in the p rivate sector to provide lifetime employment for their workers, there are various laws designed to protect workers in case of termination. In 1988 the federal government universalized the FGTS ( Fundo de Garantia de Tempo de ServiÃ§o ) program mandating emplo yers to provide severance pay to employees upon separation from employment. According to FGTS protocol, employers must pay at least 8% of the total accumulated earnings to a terminated worker regardless of the circumstances in which s/he left the company. In cases where the employee is terminated through no fault of his/her own this amount increases by an additional 40% (Martins, 2014) . Other than requiring severance pay the government also provides unemployment benefits to qualifying individuals and social security benefits upon retirement age through a system reminiscent to that used in the US and other developed nations (Balestro, 2011; Balestro, 2011) . While these programs provide a s mall measure of protection for the 16 . Among the benefits most cite healthcare, unemployment benefits, and retirement benefits.
52 Brazilian working class, scholars are quick to point out the overall growth of insecure jobs, weakening of labor unions, and reduction of earnings due to restructuring in Brazil (Bertelli, 2014) Scherer primarily attributes the growth of unemployment and sub employment in Manaus to the adoption of new capital intensive technologies by firms and the outsourcing of supportive tasks and services to external businesses. 17 She explains how increased a utomation on the assembly line reduces the need for manual workers, as firms replace relatively expensive labor with less expensive capital. She also explains how outsourcing leads to lay offs as internal support staff in services like maintenance, transpo rt, and tech support are replaced by external firms. Although she makes a good case for how outsourcing leads to job loss for industrial district employees, she fails to explore what it means in the long run for job quality. Fortunately, in 1998 Suframa b egan providing figures on the number of temporary and subcontracted employees in the Manaus Industrial District. Since 1998 such positions have fluctuated between 7 and 12 percent of all jobs at the district, with significant variation depending on the sea son. Like firms that have undergone restructuring in the US and elsewhere, businesses in the Manaus Industrial District primarily use flexible forms of labor to meet fluctuations in demand for their products. Through subcontracting and contracting temporar y employees, manufacturing firms are able to increase production during peak demand times of the year and quickly reduce their workforce thereafter. Suframa data reveals that thousands of flexible workers are hired generally between July and October and th en let go upon the conclusion of their 17 . Not to be confused with subcontracting or subcontracted employment, sub employment refers to individuals who are employed in positi ons below their qualifications and earning potential.
53 contracts at the end of the year. Between December and January of 2001 to 2002, for example, 45% of temporary workers in the district lost their jobs along with 7% of subcontracted workers; in contrast, only 3% of ful l time workers with traditional contracts lost their employment (Suframa, 2010). While this practice undoubtedly increases the efficiency of firms as they hire additional workers to fulfill short term contracts for only a few weeks or months, these positio ns are inherently less stable than traditional employment. Another University of Amazonas affiliate who has extensively studied work and organization at the MFZ, is organizational behavior theorist, Izabel Valle (2007). She examines how factories in the M anaus Industrial District modified their assembly processes, and employment structures during the 1990s in order to remain competitive in a rapidly changing market. Unlike Scherer, Valle does not focus on the adverse effects of these changes on worker live lihoods and human development. She examines how restructuring alters long established patterns and hierarchies of employment in the electronics and vehicle subsectors. Also different from Scherer, who disparages such changes, Valle suggests that the adopt ion of more capital intensive processes and increased automation can benefit the workforce. Because Valle associates capital intensive manufacturing with skilled jobs, she predicts that restructuring will increases the demand for skilled workers, such as m achine operators, in the Manaus Industrial District. Because such positions are generally better paid and more stable than unskilled manually intensive jobs, she foresees that technological advances will bring better employment opportunities for qualified individuals.
54 District have important implications for the female workforce. Is put forth in the literature skilled positions in manufacturing tend to favor male workers (Caraway, 2007; Humphrey, 1987) . Although the fact that females are overrepresented in the electronics sector in manually intensive positions is well documented, there is a current lacuna in the literature with regards t o female participation in skilled positions whether in electronics or other subsectors. With the rise of female educational attainment, and the rising incidence of workforce participation among women in Brazil, it is likely that an increasing number of wom en will benefit with the expansion of skilled employment in the Manaus Industrial District. Labor Unions and the MFZ To understand better the phenomena of flexible labor and female labor force participation in the Manaus Industrial District, it is necessar y to address the historical role of organized labor in Brazil and in Manaus. Organized labor in Brazil has a rich history dating back to the early 20 th century (Roett, 2010) . Although union organizers at times faced violent opposition from resistant emplo yers and rightwing government administrations, many unions survived the dictatorship and are active to this day. The metalworkers union in SÃ£o Paulo was a automotive indu stry. In 1953, they held a historic strike that paralyzed the industry and culminated in the release of several incarcerated union leaders. 18 Thereafter until the 18 . During the 1930s and 1940s the Vargas Dictatorship was hostile to organized labor, and manifestations were frequently met with violence. In 1953 the union held a historic strike for better salaries , which resulted in the imprisonment of several officials. After a week of negotiations however, the employers and the state met the demands of the union were and all but one of the
55 mid sixties, unions were relatively free from persecution by federal and local governments, ho wever that changed with the onset of the military dictatorship in 1964. Generally suspicious of organized labor and other leftist groups, the state dismantled Union of SÃ£o Paulo remained in a relatively subdued state until the late 1970s with the loosening of government restrictions. Reinvigorated, union members engaged in a number of historical strikes between 1978 and 1979 ( (Sindicato dos MetalÃºrgicos de SÃ£o Paulo e Mogi das Cruzes, 2009; Roett, 2010) . During the 1980s, while most of the world was undergoing restructuring, labor unions in Brazil flourished (Leite, 1997) . 19 Between 1985 and 1988 several notable strikes occurred in the Manaus Industrial District. Female union leaders and members played particularly important roles in these demonstrations, and according to one of the male leaders, females were responsible for 2005:203). It is curious that while unions in Brazil have historically been male dominated institutions, women figured prominently in the Amazonas Metalworkers Union in Manaus. 20 Anthropologist Iraildes Torres (2005) att ributes this to that fact that women have traditionally made up such a large portion of the workforce in the industrial district. She acknowledges, however, that despite the active participation of females in union imprisoned leaders were released (Sindicato dos Me talÃºrgicos de SÃ£o Paulo e Mogi das Cruzes, 2009) . 19 . According to Sociologist, Marcia Leite, while most countries in the world saw a general weakening of organized labor, accompanied by declining work conditions and decreased job stability during the 1980s organized labor in Brazil experienced a renaissance (Leite , 1997). 20 . strial labor unions rather than other unions devoted to the professions or extractive industries .
56 activities and leadership roles, women in union hierarchy frequently faced sexist discrimination from their male colleagues. Female participation in the Metal Workers Union of Amazonas is particularly interesting when one considers the literature arguing that labor unions are generally detriment al to female participation in industry. In her quantitative study of gender and factory workers from various countries, economist Terri Caraway (2007) found that wage inequality between men and women in the Brazilian manufacturing sector was more pronounce d than in East Asian countries, and there was a lower participation rate of women in industry in Brazil. She attributed this fact to the legacy of strong labor unions in Brazil, arguing that strong unions tend to favor male employment (2007). Unfortunately , a comparative study has yet to be done examining female labor union participation in Manaus compared to another region in Brazil. Such a study could further Unions continue to pl Manaus Industrial District, even though many workers themselves may not recognize it. Workers, who often criticize the effectiveness of unions, tend to argue that base wages and salaries are bel sufficiently committed to helping their constituents (DesprÃ©s, 1991; Interview, 2 2013). However, a careful look at the situation of labor in the Manaus industrial District reveals a more nua nced interpretation. Despite all the tax incentives the government extends to manufacturers in Manaus, many companies still find the location too expensive to conduct their daily operations. Shipping and labor are among the cost factors making production i n Manaus relatively expensive, thus for the sake of attracting businesses
57 and keeping them in Manaus it is critical that labor costs remain as low as possible. Through his interviews with factory workers and employers in 1984, DesprÃ©s found that while the Metalworkers Union negotiated wage increases annually, these raises were minimal. He also found that as many as 45% of those he interviewed were unaware of any action on the part of the union on their behalf and 36% claimed that they derived no value whats oever from the union (36). Worth noting, DesprÃ©s conducted his study right before a historic rash of strikes at the district; thus had he followed up with these same workers years later perhaps he would have gathered different responses. The Amazonas Meta lworkers Union has achieved a number of notable improvements and protections for workers in the decades following restructuring. Among these is the right to childcare requiring employers to provide childcare or preschool free of charge for workers with chi ldren (do EspÃrito Santo, 2013). The right to reproductive work burden generally imposed on females by the patriarchal society in which they live. Although laws requiring the s tate to provide childcare services have been in effect since 1996, demand has historically exceeded resources, thus obligating firms to fill in the vacuum left by the state. 21 Conclusion Like "national security" in the United States, "development" in Brazi l is a politically charged and nuanced subject. In recent years the political discourse promoting the benefits of the MFZ has moved away from ideas of national sovereignty, 21 . Law n. 9,394, (20t h of December 1996) Lei de Diretrizes e Bases da EducaÃ§Ã£o Nacional (Law of Directives and Bases for National Education). Article 4 section IV: Free childcare and preschool services for children from 0 to 6 years old.
58 economic integration, and regional development, to the elevation of the MFZ, and it s accompanying industrial district, as a force for preserving the rainforest through urban representatives of the European Union at EU Brazil Summit, where amid the criticisms of for the rainforest (Bizzotto, 2014) . The population of Manaus is so dependent on its free trade zone that a discontinuation of the model is no longer po litically viable. Also any decrease in the tax incentives enjoyed by the MFZ would have a severe impact on production in the industrial district, resulting in mass unemployment. The vulnerability of the model to changes in the federal tax code was exemplif ied in the early 1990s with the opening of the Brazilian economy, and the subsequent layoff of nearly half the workforce at the district. Although business in the Manaus Industrial District is currently booming, with record revenues and job numbers, criti cs are skeptical of the net positive effects on human development (Fonseca, 2013) . This is because the majority of jobs at the industrial district remain unskilled assembly positions, characterized by low wages, high turnover r ates and high incidences of injury (Moraes, 2008) . Additionally unemployment sector ( Table 2 2 ). This indicates that the model has performed below expectations in terms of quality job creation. research. Through analyses of Brazilian National census data, and qualitative interviews
59 with factory workers I seek to elucidate changes in the workforce in response to restructuring. While various scholars address gender issues in the Manaus Industrial District few studies examine how gender equality has changed over the years, particularly afte r restructuring. While various scholars lament the flexible labor policies adopted by businesses at the district and the growth of temporary and subcontracted positions with restructuring, few researchers examine the actual difference in job quality betwee n workers of different contract types, or the difference between the workers themselves.
60 Table 2 1. Mean e arnings by m anufacturing s ubsector a c omparison of MID and SÃ£o Paulo  ( c urrency v alues BRL$) Subsector Mean Monthly Earnings MID Earnings as % of SÃ£o Paulo Earnings SÃ£o Paulo MID Electronics $3,107.00 $1,539.76 50% Vehicles $3,795.00 $2,313.52 61% Metallurgy $3,129.00 $1,510.65 48% Plastics $2,199.00 $1,439.60 65% Appliances $3,189.00 $1,623.41 51% Textiles $1,826.00 $806.64 44% Source: Adapted from * Suframa. (201 4 ). Indicadores, de Desempenho do Polo Industrial de Manaus 0' 88 2010 (Page 86 ). Manaus: SuperintendÃªncia da Zona Franca de Manaus. * FIESP. (2012). Capital Humano Departamento de AÃ§Ã£o Regional: Estado de SÃ£o Paulo MÃ©dio Salarial por Setor ( t able generated on webs ite ) . Retrieved from FIESP: http://apps.fiesp.com.br/regional/dadossocioeconomicos/informacoessetor.aspx?t=3 Table 2 2. Unemployment r ates of Brazil's 20 m ost p opulous c ities (2010) City Population Unemployment Rate SÃ£o Paulo 11,244,369.00 7.5% 2.3% Rio de Janeiro 6,323,037.00 7.3% 2.1% Salvador 2,676,606.00 13.1% 7.9% BrasÃlia 2,562,963.00 8.2% 3.0% Fortaleza 2,447,409.00 7.7% 2.5% Belo Horizonte 2,375,444.00 6.5% 1.3% Manaus 1,802,525.00 11.1% 5.9% Curitiba 1,746,896.00 4.9% 0.3% Recife 1,536,934.00 12.5% 7.3% Porto Alegre 1,409,939.00 5.5% 0.3% BelÃ©m 1,392,031.00 10.3% 5.1% GoiÃ¢nia 1,301,892.00 5.3% 0.1% Guarulhos 1,222,357.00 10.0% 4.8% Campinas 1,080,999.00 6.5% 1.3% SÃ£o LuÃs 1,011,943.00 12.0% 6.8% SÃ£o GonÃ§alo 999,901.00 10.1% 4.9% MaceiÃ³ 932,608.00 12.3% 7.1% Duque de Caxias 855,046.00 11.0% 5.8% Teresina 814,439.00 9.8% 4.6% Natal 803,811.00 10.2% 5.0% Source: Adapted from* Lima, D. a. (2012). Mapa do Desemprego no Brasil . SÃ£o Paulo: AgÃªncia do Globo. Retrieved February 15, 2014, from http://oglobo.globo.com/termos de uso/
61 CHAPTER 3 CHANGING TRENDS WITH REGARDS TO GENDER OF THE WORKFO R CE Between 1990 and 1992, 34,000 people lost their jobs in the Manaus Industrial District ( Figure 3 1 ). This corresponded to the opening of the Brazilian economy under the Collor administration and the slackening of protectionist policies that had largely characterized Brazilian trade politics up to that time (Roett, 2011). The individuals who were laid off during those years were among the first casualties of restructuring in the district (Scherer, 2004). Much of the literature t hat addresses the effects of this global phenomenon on labor markets characterizes restructuring as leading to declining job quality for traditional industrial workers. Some scholars, however, have pointed out that for skilled workers, restructuring can le ad to more job opportunities. Valle for example, suggests that restructuring has led to the replacement of unskilled, manually intensive jobs with more skilled and technical positions at the Manaus Industrial District (Valle, 2007). The substitution of un skilled manual jobs with more technical positions has important implications for the female workforce. The literature on women in industry suggests that females are less likely to occupy such jobs than are males. Feminist economist, Ruth Pearson, found thi s to be case along the US/Mexican border where she observed the effects of restructuring in the early 1980s. She found that across most manufacturing subsectors the proportion of female workers declined relative to males. She attributed this decline in the proportion of positions occupied by women, to a reduction in the overall number of unskilled jobs traditionally occupied by them, as well as the growth of skilled jobs favoring male workers (Pearson, 1990). Various other
62 scholars have also suggested that restructuring leads to less gender equality in the workforce and increased poverty rates among women (Moghadam, 1999) . It is in the context of this literature that I have framed my own research on the changing trends in the wo rkforce in the Manaus Industrial District. Deviating slightly from the main conclusions put forth in the literature, however, I suggest that in the light of restructuring, women may have gained more employment opportunities relative to males. Thus, in the decades after restructuring, I expect to find a more equal distribution of the workforce by gender across various subsectors and skill levels, and less disparity in wages between males and females. There are a number of reasons to expect that women make u p a larger proportion of the industrial workforce in Manaus today than they did in decades past. For one thing, the average educational attainment of women in Brazil currently surpasses that of males. 1 With the growth of skilled and technical jobs in indus try, it Also worth noting, women have historically been an active force within the Amazonas Metal Workers Union in Manaus, frequently holding leadership positions and orga nizing collective actions. It seems appropriate that with a legacy of women in union leadership roles, these same unions are less likely to champion accords that benefit male workers at the expense of female workers. While it is not my intention to ascer tain a causal relationship between restructuring and increased employment opportunities for women, I do wish to quantify 1 . Various scholars describe education al attainment as a threshold variable for employment in the certain level, while advanced education, such as a bachelors or graduate degree may actually hinders
63 specific changes in the distribution of the Manaus industrial workforce by gender over a span of several decades. 2 Thus what follows is a presentation of original research on industrial employment in the manufacturing sector of Manaus from 1970 to 2010. By analyzing employment data from different census years I seek to answer the question, 'how has female industrial employment in Manaus f aired in light of restructuring? Although various scholars have addressed issues related to gender in factory employment at the Manaus Free Trade Zone, there is a lack of quantitative research examining changes in the composition of the workforce over tim e. This is surprising has been to fuel local development through employment. Arguably much can be learned about the economic and human development of a location through careful examination of its resident workforce. Such information is especially useful to researchers concerned with questions related to economic equality and empowerment, since increasing access to good paying jobs is among the best means to combat inequal ity. Hypothesis The hypothesis guiding this research is, that in spite of an initial decline in the female composition of the labor force in the early 1990s, in recent years women make up a larger proportion of the industrial workforce than in the years prior to restructuring. Moreover, I expect to find that in recent years (2010) a growing percentage of women 2 . Hi storÃa , they identify the period of 1991 ( Suframa/Historia ). Both of these concepts are tied to restructuring.
64 are successfully repositioning themselves within the industrial workforce in skilled positions. 3 The rest of Chapter 3 is organized in the followi ng manner. I begin by describing the methodology used in the study. I define the terms used in the study, and then describe the data set, selection of the population of interest, and the various statistical tools used in the analysis. The next section is d evoted to the research findings. It is divided into subsections that address educational attainment in Manaus, the gendered distribution of the industrial workforce over the time period in question, the distribution of the skilled industrial workforce by g ender, and differences in earnings between male and female workers. I conclude by reassessing the hypothesis in the context of the main findings, discussing weaknesses in the findings, and suggesting future research on the issue. Methods This study relies on quantitative methods to assess changes in the gendered distribution of the industrial workforce in Manaus. Few previous studies use quantitative methods. This may be partly due to disciplinary bias since many of the scholars who have examined gender is sues in the Manaus Industrial District are concentrated within the discipline of anthropology. 4 While these studies shed light on inequalities in the industrial environment and discrimination based on sex, few provide concise quantitative figures which all ow assessing changes in equality over time. 3 . occupied by workers who possess experience and professional qualifications beyond what is generally deemed as appropriate for entry level work. 4 . The vast ma jority of studies addressing the workforce at the Manaus Free Trade Zone are conducted by students and affiliates of the University of Amazonas located in Manaus.
65 Terminology While the scholarship on industrial labor distinguishes between unskilled, semi skilled and skilled job types, for the purposes of this study individuals working in semi skilled and skilled positions are combined under the single category of skilled workers. This is because the difference between skilled and unskilled positions can be subjective depending on the company one works for and in which industrial subsector one is jobs that require special training and certification, i.e. machine operators, engineers and mechanics, for the purposes of this study, the category also includes jobs in leadership and administr ative positions. Thus an individual employed as a low level supervisor of an assembly area would be classified as a skilled worker as opposed to most of the workers s/he oversees who are defined as unskilled. The Data The Brazilian National Census is cond ucted every ten years by the Brazilian Institute of Geography and Statistics (IBGE). Thousands of census agents, across the country, directly administer the questionnaires to a large percentage of the population. While the census accounts for essentially a ll households within the country, individual data are only collected for 5 to 15 percent of the population. As is generally done with large population censuses, each individual case is assigned a multiplier, indicative of the percentage of the population t hat it is expected to represent. Thus through weighting the original sample the resulting data set is representative of the entire population of the country.
66 Since the 1960s there have been various updates to the questionnaires, resulting in some diverge nces in employment and income data between different census years. 5 Nonetheless there is a lot of consistency across census years in terms of employment and occupational categories. This greatly facilitates the present research since I focus on changes in the labor force across specific industrial subsectors and job types. While IBGE conducts a number of other censuses and social surveys specifically related to employment and industry, this data is not useful for the purposes of this study, either because these focus on urban areas other than Manaus, or they fail to collect data disaggregated by gender. Selecting the Population of Study An essential step in conducting this research was limiting the data set to the select population of interest. Because th is study focuses on the workforce at the Manaus Industrial District, it was necessary to select only cases of individuals residing within the city of Manaus. Unlike many sprawling urban areas in the United States, the population of Manaus is heavily concen trated within the actual city. Because there are no adjacent urban centers within a geographic distance that reasonably allow for daily commuters, one can assume that virtually all industrial employment in Manaus is carried out by people that actually live there. In order to narrow down the census data to only the population of workers employed at the Manaus Industrial District, it was necessary to select the economically active population (EAP). While definitions of what constitutes the EAP vary slightly 5 . One example of this is how only after 1980 did IBGE begin collecting information on wheth er or not employed individuals had a work card signed by their employer. Another example is how prior to 1991, data was not collected on whether or not individuals had worked during the previous week (Jannuzzi, 2003) .
67 depending on the research institution, this study follows the guidelines set forth by IBGE on its official website (IBGE, 2014) . The EAP consists of all individuals 15 years of age or older who either work or are looking for work. This includes parts of the population that work or volunteer 15 or more hours a week without receiving pay. Not considered part of the EAP are students, stay at home partners, retired persons and persons that otherwise live on pensions. 6 Because thi s study focuses specifically on the workforce at the Manaus Industrial District, only cases representative of manufacturing activities that take place at the Manaus Industrial District have been selected. Excluded from the analysis are subsectors that rese mble cottage industries, such as small scale garment manufacturers, makers of customized clothing, accessories and artisanal items. Additionally, any cases with a vaguely defined industrial activity have been excluded from the analysis. 7 By selecting only cases that fall within the twenty two industrial subsectors, monitored by Suframa, the pool of cases best approximates the actual population of industrial workers employed at the Manaus Industrial District. 8 It is worth mentioning that there are inconsist encies between the employment numbers represented in official Suframa reports and the number of cases selected in this study. This is likely due to the fact that, depending on the census year, individual level data was collected for only 6% to 17% percent of the actual population of Manaus. 6 . Implic it in these categories of individuals are retired people, and individuals who live on social security. 7 . 8 . Companies are generally grouped by Suframa in one of twenty two subsectors based on the primary products they produce. The electronics subsector and the vehicles subsector are historically the two most prominent sectors in terms of revenue and employment generation.
68 It is also worth noting that, between as many as 15% to 20% of these cases have incomplete employment data, i.e. the census agent noted an individual's occupational category, but failed to specify in which industrial sec tor they were employed. Since both of these variables are critical to the analysis of the change in the gendered distribution of the workforce at the Manaus Industrial District, any cases that are incomplete in this manner were excluded. The end result is a weighted sample that is roughly 15% to 25% smaller than the actual population of interest based on Suframa's published report. Because this study focuses on labor force participation rates by sex and the gender composition of the workforce, rather than o n the absolute growth of employment in Manaus, it is not problematic that the total number of cases included in the analysis is inferior to the actual number of workers reported by Suframa. The data published by Suframa reflect the changes in industrial em ployment in Manaus since 1988. Unfortunately however, Suframa only began disaggregating employment data by gender as recently as 2004 ( Table 3 1 ). This does not allow one to examine changes with regards to the gendered composit ion of the workforce before and after restructuring in the 1990s, which is essential to addressing the hypothesis, in question (Suframa, 2010:129). Additionally Suframa does not disaggregate skilled and unskilled workers, nor do they provide figures on act ual wages paid to the workforce. 9 Tools of Analysis I used the statistical software package SPSS, to disaggregate the census data and perform a number of statistical tests. Because the research question centers on the 9 . Suframa does not provide data with statistic s like median or standard deviation, which indicate actual wages paid out to the population of industrial workers. They do list the total average monthly cost of wages to each subsector, which is divided by the number of total employees listed in each sect or. The quotient can be roughly interpreted as the average monthly earnings for each employ in the subsector with no account to actual distribution of within the subsector.
69 distribution of men and women in the i ndustrial workforce, the bulk of the analysis was tabulations feature. I also used the means comparison feature when analyzing income data. I initially treat the census years as separate data sets, analyzing each one indi vidually then aggregating the findings into tables for comparison. Recoding An important step in the analytical process was reviewing the industrial correspond to the subsectors of the Manaus Industrial District. Census takers collect fairly specific employment information for individual cases, which can be used to disaggregate the weighted sample by industry and occupational category. Any cases of industries that fell outside the 21 industrial subsectors specified by Suframa were excluded from the analysis, i.e. agricultural and extractive industries, the public sector (including education), the service sector, etc. In order to assess the distribution of men and women a mong skilled positions, it was necessary to create a binary variable corresponding to skilled and unskilled workers. The data could then be disaggregated by gender and the new variable, showing the distribution of the sexes among skilled and unskilled posi tions as well as the respective indices of men and women in such positions. When recoding occupation type into the binary variable, skilled worker, it was necessary to consider hundreds of different occupational categories, and determine whether each shou ld be classified as skilled or unskilled. 10 Fortunately for purposes of 10 . I considered both the literature on industrial employment and knowledge gained through qualitative fieldwork when determining which occupational categories were skilled and unskilled (Humphrey
70 comparison, there is a high degree of consistency between census years in terms of occupation categories. The following job categories, for example, were assigned the oded into the binary variable skilled workers, across all of the census years: mechanics, technicians, managers, supervisors, welders, quality controllers, administrators and engineers. Other occupational categories representative of unskilled workers, i.e . general assembly workers, telephone receptionists, janitors etc. were assigned a value of 0. 11 Means Comparison I primarily used the means comparison feature in SPSS to assess differences between the earnings of males and females employed within the 21 subsectors representative of the Manaus Industrial District. I consider the difference between men ings in terms of both median and mean wages, expressing female earnings as a percentage of male earnings. The reason for expressing female earnings as a percentage of male earnings is to facilitate comparison across census years since there were substantia l changes in the Brazilian national currency and minimum wage laws during the 40 years represented by the census data. Even after converting the monetary unit in each census year to the equivalent in Brazilian reais for the year 2010, comparisons in terms of real amounts was impractical. 1987; Jannuzzi, 2003). Although this process was somewhat subjective it is important to remember that a high degree of consistency was maintai ned across census years. In other words all occupations I recoded as skilled in 1970 are considered skilled in 1980 and subsequent years. 11 . It should be noted that there were differences in the types occupations and the total number of occupational catego ries represented in different census years. Especially after 1991 additional luded it among skilled jobs for that and subsequent years.
71 A comparison of average and median earnings between male and female industrial workers is not only an important method of assessing gender equality, but also reveals a great deal about occupational segregation. As put forth in the literature most differences in earnings between male and females can be attributed to occupational segregation, whereby men on average occupy better paying positions than do women (Chant, 2003). 12 Thus a decline in wage gaps implicitly indicates a m ore equal distribution of males and females among occupational categories. Findings Economically Active Population Between 1970 and 2010 the population of Manaus increased substantially. In 1970, only two years after the Manaus Industrial District opened, Manaus had little more than 300,000 inhabitants. By 1980 the population had more than doubled, increasing to 633,383. By the following decade (1991) the population surpassed one million. Between 1991 and 2000 the population increased by another 400,000 pe ople, and then 400,000 more over the next decade. By 2010 the total population of Manaus numbered 1.8 million, an increase of nearly 600% since 1970, which is a testament to the migration attracted by the industrial district ( T able 3 2 ). During the same period the incidence of economic activity generally increased. In 1970 only 50% of the population of Manaus was economically active, with participation rates of 80% for men and 24% for women, respectively. By 1980 the 12 . The 1988 Brazilian Constitution expressly prohibits wage discrimination based on sex. Although there may be isolated cases where employers in the Industrial District defy the law, and pay lower wages to their female employees than to males occupying the same positions, this is unlikely to account for much of the difference observed between men and women in median and average wages. It is far more likely that any difference in mean and median wage s is due to segregation by subsector and occupation, with women typically occupying lower status positions in subsectors that pay less on average.
72 incidence of economic activity among the population increased to 58%, with participation rates of 81% and 37% for males and females, respectively. In 1991, however, 53% of the population of Manaus was economically active, a decrease of 9% from the previous decade. T he respective participation rates of men and women among the economically active population decreased to 72% and 36%, during that same period. By 2000 the percentage of the population counted as economically active reached its highest point of the years st udied, at 65% of the population. During that same year the female economically active participation rate increased to 52% whereas for males it was only 78%. These figures changed little over the next decade, with 64% of the population of Manaus classified as economically active in 2010, with respective incidences of activity for males and females of 75% and 55% ( Table 3 2 ). It is important to consider these figures when studying gender equality in terms of the distribution of t he workforce in the Manaus Industrial District. This is because the percentage of women in the initial labor pool has important implications on the gendered distribution of the workforce in the Manaus Industrial District. Educational Attainment My finding s regarding differences in educational attainment between men and women in Manaus are consistent with most of the literature on education in Brazil. 13 Educational levels have risen significantly since 1970 and women have generally outpaced men in terms of e ducational attainment. Since 1970 the share of high school 13 . The literature on educational attainment in Brazil indicates that educational levels have risen significa ntly since 1970 and that females have generally outpaced males in terms of educational attainment (Nonato et al . 2012).
73 has increased. As displayed in Table 3 3 , 23% of women and 12% of men in 1970 were high s chool graduates. 14 Although the percentage of high school graduates among the female economically active population stagnated between 1970 and 1980, the percentage of male graduates increased to 16%. In the following census year (1991), the respective incid ences of high school graduates among the male and female EAP were 29% and 38%. In 2000, the incidence of high school graduates increased to 33% for men and 42% for women. Finally by 2010, 51% of men and 60% of women in the EAP had graduated from high schoo l, indicating a majority of high school graduates among the economically active population. The incidence of high school graduates among the population of workers employed in the 21 manufacturing subsectors representative of the Manaus Industrial District follow a trend similar to that of the general economically active population. In 1970 only 8% of men and 8% of women, employed in the 21 subsectors, were high school graduates ( Table 3 4 ). By 1980 the incidence of high school graduates among the male and female workforce were 14% and 10%, respectively. 15 The following decade the incidence of high school graduates among the female workforce had more than tripled to 35% and also increased substantially for males, to 31%. This la rge increase in the incidence of high school graduates between 1980 and 1991 may be attributed to the fact that more and more firms began to require that their assembly 14 . Included among high school graduates are individuals who have attended some college but have not yet graduated. 15 . Both in 1970 and 1980 the incidence of high school graduates among the workforce of the Manaus Industrial District was significantly lower than the economically active population of Manaus in general. While there could be a number of reasons for this, one might be that the majority of the workforce at the district was made up by migrants from the countryside, who had less access to education prior to locating in Manaus (DesprÃ©s 1991).
74 workers be high school graduates. By 2000 the incidence of high school graduates among the manufacturing workforce had increased to 46% for men and 66% for women, surpassing that of the general economically active population. This trend continued to 2010 wherein 84% of female workers and 70% of male workers were high school graduates, such t hat the incidence of high school graduates was much higher than that of the general EAP. The incidence of college graduates among the economically active population in Manaus, and more specifically, the workforce of the Manaus Industrial District (MID) fol low trends similar to those with regards to high school graduates. One notable difference, however, is that among the economically active population, the incidence of college graduates among women began to exceed that of men by 1980, whereas within the man ufacturing workforce, women did not surpass men in this respect until 2000. By 2010, 13% of women employed in the Manaus Industrial District were college graduates compared to only 7% of men. It is also worth noting that the incidence of college graduates among the MID in 2010 was lower than among the general economically active population. When one considers how education has been described as a reduces the likelihood of empl oyment in the manufacturing sector, this is less surprising (Caraway, 2007) . The Manufacturing Sector While the bulk of my analysis on employment in Manaus is limited to just the 21 subsectors of the Manaus Industrial District, here I briefly consider all manufacturing subsectors together. I do this in order to observe changes in the distribution of the total working economically active population across employment sectors and to determine
75 changes with regards to the relevancy o f manufacturing in Manaus in the context of historical macro economic trends. 16 In 1970, when the Manaus Free Trade Zone was still in its infancy, only about 13% of the working economically active population was employed in manufacturing. By 1980 this figu re had increased to more than 28%. The rapid growth in that decade is indicative of the maturation of the Manaus Industrial District, which went on to flourish during the 1980s. However, by the following decade, in light of the opening of the national econ manufacturing decreased to only 22% of the occupied EAP. By 2000, manufacturing continued to lose relevancy as an employer, as only 13% of jobs were reported to be in manufacturing. 17 Altho ugh the distribution of the working population among manufacturing jobs had increased to 16% by 2010, this remained below levels prior to restructuring. Gendered Distribution of the Workforce in the Manaus Industrial District During this same time there we re substantial changes in the gendered distribution of the workforce of the MID. In 1970 women constituted 19% of the workforce in the 21 16 . By selecting for only the 22 manufacturing subsectors representative of the Manaus Ind ustrial District, one achieves a weighted sample that is smaller than the actual population of interest due to missing cases where specific employment data for individuals were not collected. While the weighted sample is significant for the population of w orkers employed at the MID, the total number of cases does not reflect the actual number of people there employed. This is substantiated through comparing official employment data from Suframa with census data. It would therefore be problematic to consider only the 22 subsectors representative of the MID when examining the distribution of the total working population by employment subsectors. 17 . in years pas f convenience, rather than for lack of a more appropriate classification. This assumption is based on the fact that just 9 years earlier, distribution of t he workforce across employment sectors in 2000 and 2010, it is important to recognize that workforce distribution is probably underrepresented in at least some of the categories.
76 subsectors representative of the industrial district. Although women accounted for nearly half of the workforce in ele ctronics at that time, only 1% of the jobs were in that subsector. The subsector of textiles, which also employed a lot of women, was the most prominent employer at the MID, accounting for 29% of total jobs ( Table 3 5 ). By 1980 , with rapid expansion of the electronics subsector, women occupied 39% of the jobs in the district, an increase of more than 100% from the previous decade. At that time women formed the majority of workers within the following subsectors: electronics (56% ), timepieces (53%), garments (68%), and textiles (53%). Also worth noting, in 1980 women made up 10% of those employed in vehicle manufacturing, a subsector that remains largely male dominated throughout the rest of the census years ( Table 3 5 ). In 1991 women made up 42% of the workforce of the Manaus Industrial District. When one considers that the female economically active participation rate was only 36% compared to 72% for males during that same year, one begins to comprehend the role of the Manaus Industrial District as an employer of women. While women continued to make up a narrow majority of those employed in the electronics subsector (51%), they also made up a large majority of the workers (91%) employed in the garments subsector ( Table 3 5 ). After 1991 there is a trend towards defeminization of the industrial district. 18 By 2000, only 33% of the MID workforce was female. Because Suframa has published total employment figures for the years 1991 and 2000, one can observe that the Manaus 18 . formerly employed substantial numbers of female workers, began to shift toward hiring more males and less female employees.
77 Industrial District actually employed fewer workers in 2000 than in the previous decade. By the following decade, however, the process of defeminization of the industrial district appears to have stagna ted. According to the census data, women made up 32% of those employed in the MID in 2010. Fortunately, Suframa has also provided data regarding the gendered distribution of the workforce during that year, indicating that in July, 2010, 29.4% of the workfo rce employed at the industrial district was female. A difference of merely 3% between the census data and figures published by Suframa suggests that the census data fairly accurately capture the reality of the gender distribution of the workforce in the Ma naus Industrial District. Skilled Jobs Despite the generalized defeminization of the manufacturing workforce in Manaus between 1991 and 2000, the share of women in skilled positions actually improved in certain subsectors. These findings have important i mplications for gender equality in the Manaus Industrial District because they indicate that women gained access to better paying, potentially more stable employment, and that the traditional gendered division of factory labor in Manaus may be in decline. It is worth noting that the share of jobs that were skilled in the Manaus Industrial District generally improved between 1970 and 2000. As shown in Figure 3 2 in 1970 only 24% of all the jobs in the industrial district were co nsidered skilled by the year 2000, 51% of jobs were skilled. In 2010, however, subsequent to the rapid growth of the total workforce, only 39% of the positions in the MID were skilled. Although we are more concerned with changes in the gendered distributio n of the skilled workforce than changes in the overall composition of the workforce in terms of skilled jobs, the latter has important implications for gender equality as women are generally
78 underrepresented in skilled positions, and skilled jobs tend to p ay better than unskilled positions (Jannuzzi, 2003). The share of women among skilled positions in the MID significantly increased during the first decade examined in this study thereafter plateauing at roughly 26%. In 1970 women only occupied 9% of skille d positions in the industrial district; by 1980 the share of females in such positions had increased to 25%. 19 In the following decades (1991 2010) the share of women in skilled positions in the MID remained at 26% ( Table 3 6 ). Although the share of skilled jobs occupied by women remained at 26% between 1991 and 2010, it changed substantially in the electronics subsector during that period. Whereas the female share of skilled positions within the electronics subsector in 1991 was only 32%, the ratio slightly increased in favor of women to 34% in 2000. By 2010 it had reached 48%, with women occupying almost half of all skilled positions in the subsector ( Table 3 6 ). These findings suggest that after th e opening of the Brazilian economy and the onset of restructuring women in skilled positions were less affected than were those in unskilled positions. While there was a general defeminization of unskilled jobs between 1991 and 2000 at the MID, that does n ot appear to have been the case for skilled workers. Differences in Earnings between Men and Women The difference in mean earnings for men and women generally improved over the census years examined in this study. In 1970, mean female earnings in 19 . Within electronics, females constituted a large minority, 44% of the skilled workforce, and also occupied a majority of such po sitions (57%) in the printing subsector (Table 3 6).
79 manufacturing in Manaus were only 45% of male earnings, indicating that on average, female workers at the MID earned less than half of what male workers earned ( Table 3 7 ). However during the same year, female median earnings a s expressed as a percentage of male earnings was 67% suggesting that the pronounced difference in mean earnings between men and women in 1970 may have been influenced by a relatively small number of men receiving extremely high wages. By the following dec ade (1980), the gap in mean earnings had decreased slightly with women earning on average 49% of what men earned. That same year the gap in median earnings actually increased with female median earnings equaling only 63% of male earnings. In the following decades the gap in mean earnings continued to decrease, falling to 42% in 1991 and 25% in both 2000 and 2010, respectively. Median earnings followed a similar trend during the period, with female median earnings equaling 72% of male median earnings in 1991 , 86% in 2000 and 85% in 2010. By disaggregating mean earnings by skilled and unskilled workers, we gain additional insight into the nature of the wage gap. As displayed in Figure 3 3 , the gap in mean earnings between unskille d men and women generally decreased over the census years, improving from 67% in 1970 to 94% in 2000. The wage gap increased in 2010, however, with mean female earnings declining to only 87% of male earnings that year. Although the reason for this decline in female earnings relative to male earnings is unclear, it may be related to the changing gender composition of the workforce in the various subsectors.
80 The gap in earnings between men and women in skilled positions also improved substantially over the course of the census years. In 1970 female skilled workers earned on average 51% of what male skilled workers earned. Although in 1980 mean female skilled earnings declined to 46% of male earnings, these improved in subsequent decades. The overall improvem ent in the wage gap between 1980 and 2010 was 87%, as differences in earnings progressively declined by 32%, 23%, and 13% over each decade ( Figure 3 4 were 86% those of skilled ma census years. The data for 2000 to 2010 suggest that the general wage gap between male and female workers in the MID is highly correlated with the skilled composition of the workforce in terms gend er. While the gap in earnings between skilled male and female workers declined in that decade, it increased for men and women in unskilled positions. Had there not been a reduction in the share of skilled positions during the decade, it is likely the gener al wage gap would have continued to fa ll. These findings suggest that a more equal distribution of the skilled workforce by gender would bring greater equality in terms of overall earnings between men and women. Conclusion Analysis of data from the 1970 2 010 censuses indicates that there were substantial changes in the gendered distribution of the workforce in the Manaus Industrial District during the period. These changes suggest both advancements and anufacturing workforce. One of the main setbacks with regards to gender equality was the marked decline in the female share of the industrial workforce between 1991 and 2000.
81 Although this decrease in the proportion of jobs occupied by women was not equal ly pronounced across all manufacturing subsectors, there was a general defeminization of labor at the Manaus Industrial District during the period. The female share of the workforce in the Manaus Industrial District notably declined after restructuring. A lthough I expected to observe an initial decline after 1991 basing my assumption on literature describing the phenomenon of labor defeminization, the observed decrease of 21% was greater than anticipated. I also expected to view an increase in the female s hare of the workforce between 2000 and 2010, due to continued advances in female educational attainment, the growth of the female EAP, and record employment levels in the MID in recent years. Nonetheless I observed no improvement, with females making up 33 % and 32% of the manufacturing workforce during those years, respectively. These findings contradict the initial hypothesis that the female The findings regarding differ ences in earnings between men and women in the Industrial district are indicative of progress towards greater gender equality up until 2000. While the gap in earnings between male and female workers continued to improve through 2010, the difference in earn ings for unskilled workers actually increased between 2000 and 2010 indicating a reversal of gains from the previous decade. Because income is highly contingent on job type, these findings suggest a more equitable distribution of women and men among occupa tional categories after 1980. My findings regarding the distribution of women in skilled positions suggest that the female share of the skilled workforce at the Manaus Industrial district initially
82 increased between 1970 and 1980, plateauing thereafter at 26%. This suggests that while there was a general defeminization of the MID workforce during those years, this mostly occurred among the unskilled portion of the workforce. This is further substantiated by the share of women in skilled positions which pea ked at 43% in 2000. While the share of the female workforce in the Manaus Industrial District employed in skilled positions decreased in 2010 to only 31%, this percentage is still higher than prior to restructuring ( Table 3 8 ). The distribution and incidence of women among skilled positions in the MID has important implications for gender equality in Manaus. It suggests that more women have access to relatively better paying, more secure employment. The growth of better employm ent opportunities for women in manufacturing may reflect gains made in female education levels, and the decline of cultural barriers that limited women in industry during previous decades. ugh empowerment, defined as a process whereby individuals take control of their own lives, can be difficult to quantify, financial autonomy and professional achievement are good indicators of empowerment. Thus, evidence that more women in the MID are gaini ng access to better employment, might suggest that women have more control over their own educational and professional qualifications, allowing them to land such jobs. Furthermore, relatively better wages and job security associated with skilled positions , imply greater financial autonomy and professional achievement, indicators of empowerment.
83 There are challenges inherent in using national census data, including the fact that such data are often incomplete. Despite such challenges, however, the census data used in this study served well for the task of measuring changes in the manufacturing workforce over the period of interest. Although a certain amount of caution should be applied when considering the statistics cited among these findings , since these are based on an approximation of the actual population of interest.
84 Source: Adapted from Suframa. (2010). Indicadores, de Desempenho do Polo Industrial de Manaus 1988 2010 (Page 7 1 ). Manaus: SuperintendÃªncia da Zona Franca de Manaus. Suframa. (2014). Indicadores de Desempenho do PÃ³lo Industrial de Manaus 2008 2014 (Page 96 97). Manaus: SuperintendÃªncia da Zona Franca de Manaus. Figure 3 1. Historical v ariation of the w orkforce in the Manaus Industrial District 0 20000 40000 60000 80000 100000 120000 140000 total number of workers Year total workforce female workers
85 Table 3 1. Change in w orkforce at the Manaus Industrial District 1998 2013 ( a nnual a verages) Month/Year N. Companies N. Fully contracted N. Temporary Workers TOTAL N. Subcontracted Workers Female Labor Flexible workers as % of Total Workforce Female Labor as % of Total Workforce 1998 313 45574 1531 49583 2478 8.1% 1999 305 38602 1611 43113 2901 10.5% 2000 312 43896 2956 50004 3153 12.2% 2001 327 48426 3072 54732 3234 11.5% 2002 346 51137 3443 57812 3232 11.5% 2003 354 57524 3579 64971 3868 11.5% 2004 365 70013 5418 79381 4017 26911 11.9% 33.9% 2005 397 81868 4052 89869 3950 31343 8.9% 34.9% 2006 417 89259 4859 98666 4548 31363 9.5% 31.8% 2007 411 89024 4948 98720 4749 29741 9.8% 30.1% 2008 419 96905 5641 106914 4367 31692 9.4% 29.6% 2009 417 84932 3802 78026 3739 26772 9.7% 34.3% 2010 431 71949 5820 103663 4631 29944 10.1% 28.9% 2011 448 110623 4601 119909 4685 35681 7.7% 29.8% 2012 462 111697 3920 120184 4567 37619 7.1% 31.3% 2013 472 112846 3659 121249 4744 38321 6.9% 31.6% Source: Adapted from Suframa. (2010). Indicadores, de Desempenho do Polo Industrial de Manaus 1988 2010 (Page 73). Manaus: SuperintendÃªncia da Zona Franca de Manaus. Suframa. (2014). Indicadores de Desempenho do PÃ³lo Industrial de Manaus 2008 2014 (Page 96 97). Manaus: SuperintendÃªncia da Zona Franca de Manaus.
86 Table 3 2. Total p opulation of the s tate of Amazonas and Manaus and % of p opulation that is e conomically a ctive Census Year 1970 1980 1991 2000 2010 Manaus Pop. 311,622 633,383 1,011,726 1,401,395 1,813,461 in Total Pop. (Manaus) 0% 103% 60% 39% 29% Amazonas Pop. 961,408 1,449,135 2,103,245 2,808,333 3,483,985 Pop. (Amazonas) 0% 51% 45% 34% 24% % of Population Economically Active (Manaus) 50% 58% 53% 65% 64% % of Men Economically Active (Manaus) 80% 81% 72% 78% 75% % of Women Economically Active (Manaus 24% 37% 36% 52% 55% Source: Compiled from Brazilian National Censuses (1970 2010)
87 Table 3 3. Incidence e ducational a ttainment of the e conomically a ctive p opulation of M anaus Year Gender High School Graduate* College Graduate 1970 Male 12% 4% Female 23% 5% 1980 Male 16% 4% Female 23% 5% 1991 Male 29% 6% Female 38% 7% 2000 Male 33% 4% Female 42% 5% 2010 Male 51% 10% Female 60% 15% * Includes individuals that have attended some college but never graduated. Source: Compiled from Brazilian National Demographic Censuses (1970 2010)
88 Table 3 4. Incidence e ducational a ttainment of m anufacturing w orkforce, by g ender Year Gender High School Graduate* College Graduate 1970 Male 8% 2% Female 8% --1980 Male 14% 2% Female 10% 1% 1991 Male 31% 5% Female 35% 3% 2000 Male 46% 3% Female 66% 4% 2010 Male 69% 7% Female 84% 13% * Includes individuals that have attended some college but never graduated. Source: Compiled from Brazilian National Demographic Censuses (1970 2010)
89 Table 3 5. Gendered c omposition of MID w orkforce , d isaggregated by s ubsector (1970 2010) Year Electronics Time Pieces Vehicles Metallurgy Appliances Plastics Textiles Garments Other* Total 1970 Male 53.0% 97.0% 96.0% 99.0% 95.0% 55.0% 82.0% 88.8% 81.0% Female 47.0% 3.0% 4.0% 1.0% 5.0% 45.0% 18.0% 11.2% 19.0% Subsector as % Total 1.0% 4.0% 7.0% 15.0% 1.0% 29.0% 2.0% 41.0% 100.0% 1980 Male 44.0% 47.0% 90.0% 85.0% 61.0% 65.0% 47.0% 32.0% 86.0% 61.0% Female 56.0% 53.0% 10.0% 15.0% 39.0% 35.0% 53.0% 68.0% 14.0% 39.0% Subsector as % Total 47.0% 1.0% 9.0% 3.0% 1.0% 3.0% 7.0% 1.0% 28.0% 100.0% 1991 Male 49.0% 73.0% 88.0% 85.0% 60.0% 71.0% 54.0% 9.0% 84.9% 58.0% Female 51.0% 27.0% 12.0% 15.0% 40.0% 29.0% 46.0% 91.0% 15.1% 42.0% Subsector as % Total 65.0% 1.0% 5.0% 4.0% 4.0% 4.0% 2.0% 1.0% 14.0% 100.0% 2000 Male 56.0% 30.0% 88.0% 81.0% 71.0% 78.0% 52.0% 17.0% 80.1% 67.0% Female 44.0% 70.0% 12.0% 19.0% 29.0% 22.0% 48.0% 83.0% 19.9% 33.0% Subsector as % Total 39.0% 3.0% 10.0% 4.0% 6.0% 5.0% 1.0% 1.0% 31.0% 100.0% 2010 Male 48.0% 86.0% 88.0% 71.0% 74.0% 60.0% 23.0% 74.1% 68.0% Female 52.0% 14.0% 12.0% 29.0% 26.0% 40.0% 77.0% 25.9% 32.0% Subsector as % Total 34.0% 0.0% 18.0% 10.0% 7.0% 6.0% 1.0% 1.0% 23.0% 100.0% * Includes 13 remaining subsectors: Beverages, Wood, Paper , Chemical, Cleaning, Food Items, Printing, Non metallic , Furniture, Optic, Hygiene, Naval and Diverse Compiled from Brazilian National Demographic Census (1970 2010)
90 Table 3 6. Distribution of skilled MID workforce by gender, disaggregated by subsector (1970 2010) Subsector 1970 1980 1991 2000 2010 Male Female Male Female Male Female Male Female Male Female Electronics 55.6% 44.4% 64.7% 35.3% 67.8% 32.2% 65.6% 34.4% 52.5% 47.5% Time pieces --45.7% 54.3% 83.0% 17.0% 32.7% 67.3% --Vehicles 94.7% 5.3% 92.5% 7.5% 84.2% 15.8% 91.6% 8.4% 85.9% 14.1% Metallurgy 93.5% 6.5% 92.6% 7.4% 93.1% 6.9% 90.4% 9.6% 89.6% 10.4% Appliances 98.9% 1.1% 78.8% 21.2% 80.5% 19.5% 76.8% 23.2% 87.5% 12.5% Plastics 88.9% 11.1% 88.3% 11.7% 83.8% 16.2% 80.5% 19.5% 77.5% 22.5% Textile 86.7% 13.3% 88.2% 11.8% 92.8% 7.2% 50.0% 50.0% 37.5% 62.5% Garments* 85.7% 14.3% 75.0% 25.0% 23.1% 76.9% 43.8% 56.3% --Other** 80.7% 19.3% 88.2% 11.8% 82.1% 17.9% 77.6% 22.4% 73.1% 26.9% Total 90.8% 9.2% 75.2% 24.8% 73.6% 26.4% 74.0% 26.0% 73.8% 26.2% * Includes Footwear and leather products **Includes 15 remaining subsectors: Includes 13 remaining subsectors: Beverages, Wood, Paper , Chemical, Cleaning, Food Items , Printing, Non metallic, Furniture, Optic, Hygiene, Naval, and Diverse Compiled from Brazilian National Demographic Census (1970 2010)
91 Table 3 7. Monthly e arnings of MID workforce, d isaggregated by g ender [1970 2010] ( c urrency values BRL$, a djusted to 2010) Census Year Sex Mean Monthly Earnings Mean Female Earnings as % of Mean Male Earnings Female Earnings as % of Mean Male Earnings Median Monthly Earnings Median Female Earnings as % of Median Male Earnings Median Female Earnings as % of Median Male Earnings N. Std. Dev. (Mean) 1970 Male $419.67 45.0% 0.0% $213.12 67.5% 0.0% 5509 765.8 Female $190.29 $143.86 1142 187.7 Both $380.28 $213.12 6651 706.6 1980 Male $1,295.99 49.0% 8.9% $809.65 62.5% 7.4% 20332 1838 Female $630.04 $506.03 12899 499 Both $1,037.49 $607.24 33231 1506 1991 Male $2,465.66 58.0% 18.4% $1,445.71 71.7% 14.7% 20743 3952 Female $1,433.33 $1,036.09 10148 1414 Both $2,126.52 $1,265.00 30891 3374 2000 Male $1,026.70 75.0% 29.3% $676.03 86.9% 21.3% 9485 1174 Female $773.80 $587.68 7901 620.5 Both $911.77 $652.98 17387 971.3 2010 Male $1,237.45 75.0% 0.0% $800.00 85.0% 2.2% 32706 811.6 Female $922.25 $680.00 15958 1637 Both $1,134.08 $780.00 48664 1428 Source: Compiled from Brazilian National Demographic Census (1970 2010)
92 Table 3 8. Share of MID W orkforce in s killed p ositions, d isaggregated by g ender s ubsector Sub Sector 1970 1980 1991 2000 2010 % of Male Skilled % of Female Skilled % of Male Skilled % of Female Skilled % of Male Skilled % of Female Skilled % of Male Skilled % of Female Skilled % of Male Skilled % of Female Skilled Electronics 71.4% 66.7% 41.3% 17.4% 60.8% 27.4% 51.1% 34.9% 31.0% 26.7% Time pieces 100.0% 100.0% 100.0% 100.0% 81.8% 78.7% Vehicles 23.4% 100.0% 30.3% 20.3% 37.4% 48.9% 62.1% 44.1% 40.0% 39.4% Plastics 19.5% 20.0% 38.7% 9.7% 45.8% 22.1% 80.5% 68.8% 71.0% 52.3% Metallurgy 53.7% 80.0% 57.5% 23.6% 53.4% 24.9% 79.2% 56.5% 66.0% 49.2% Appliances 78.7% 100.0% 52.5% 22.3% 70.6% 25.6% 61.0% 49.1% 40.0% 15.3% Textile 17.8% 3.9% 27.2% 3.3% 40.4% 3.5% 16.0% 22.2% 13.0% 29.4% Garments* 27.3% 50.0% 38.4% 5.5% 15.0% 5.6% 43.8% 16.4% Other** 5.3% 10.4% 5.5% 4.2% 5.8% 6.9% 14.7% 18.0% 9.5% 11.2% Total 27.0% 10.3% 33.7% 17.2% 53.1% 27.7% 54.8% 43.4% 42.0% 31.6% * Includes Footwear and leather products * Includes 15 remaining subsectors: Includes 13 remaining subsectors: Beverages, Wood, Paper , Chemical, Cleaning, Food Items , Printing, Non metallic, Furniture, Optic, Hygiene, Naval, and Diverse Compiled from Brazilian National Demographic Census (1970 2010)
93 Source: Compiled from Brazilian National Demographic Census (1970 2010) Figure 3 2 Composition of the MID w orkforce by s killed and u nskilled j obs 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1970 1980 1991 2000 2010 Year skilled workers unskilled workers
94 Source: Compiled from Brazilian National Demographic Census (1970 2010) Figure 3 3. Wage g ap between m ale and f emale u nskilled MID w orkers 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1970 1980 1991 2000 2010 Year corresponding wage gap av. female unskilled earnings % male unskilled
95 Source: Compiled from Brazilian National Demographic Census (1970 2010) Figure 3 4. Wage g ap between m ale and f emale s killed MID w orker s (1970 2010) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1970 1980 1991 2000 2010 Year corresponding wage gap av. female skilled earnings % male skilled
96 CHAPTER 4 FLEXIBLE WORKERS AND GENDER As home to Brazil's only free trade zone and the largest electronics manufacturing center in South America, the city of Manaus provides a unique context in which to study the phenomena of labor flexibilization and labor feminization (Schneider, 2005) . The literature on restructuring associates labor flexibilization with outsourcing, short term employment contracts, declining employee benefits, and decreasing job stability (BenerÃa L. , 2003) . While sc holars have written extensively on the Manaus Free Trade Zone (MFZ) and there are various studies that focus on factory employment in the industrial district, there remains a significant lacuna in the literature with regards to subcontracted and temporary workers. These types of workers, which exemplify labor flexibilization, currently constitute about 7% of the workforce of the district ( Table 3 1 ) (Suframa, 2014) . Unfortunately, Suframa does not pr ovide data on subcontracted and temporary workers disaggregated by gender, wages, and manufacturing subsector; thus we are unable to determine the varied usage of flexible workers across industries nor detect wage and gender inequality between workers base d on contract type. Today women make up about 30% of the workforce in the Manaus Industrial District. Quantitative analysis of Brazilian national census data suggests that while the share of women in manufacturing declined after the opening of the nation al economy in 1990, there was a shift to more women occupying skilled positions during that time. While these data suggest that the share of skilled jobs occupied by women generally
97 plateaued between 2000 and 2010, there was a significant shift towards fem ales within the electronics subsector. 1 Here in Chapter 4 I present the results of the qualitative field work I conducted in Manaus during the summer of 2013. A primary goal of the research was to gather porary contracts. I thus conducted 45 semi structured interviews with individuals who worked for or with firms located in the Manaus Industrial District either as flexible workers (subcontracted and temporary) or fully contracted workers (traditional full time employees). By design, about half of the research participants were female; this was to facilitate subsequent analyses regarding gender in industrial employment. By considering differences in responses among male and female participants I gained a gre ater perspective on how gender influences the opportunities and experiences available to the hardworking individuals of Manaus who make their living through industrial employment. Hypotheses Regarding Flexible Workers The assumption that subcontracted and temporary forms of employment in the Manaus Industrial District are inferior to traditional full time jobs influenced the principal research questions. The main hypothesis guiding this research on flexible workers at the Manaus Industrial District is as f ollows: 1. Flexible workers and their fully contracted peers differ in terms of educational attainment and professional qualifications. This hypothesis is based on the assumption that subcontracted and temporary positions are generally less desirable than e mployment under traditional contracts, thus 1 . As displayed in Table 3.10 between 2000 and 2010 the proportion of skilled positions occupied by females within the electronics sub sector increased from 34.4 to 47.5%.
98 less qualified individuals would tend to occupy such positions. Also, there is the assumption that firms follow a selection process for hiring flexible workers (subcontracted and temporaries) that is less rigorou s than what is followed by firms hiring individuals for fully contracted positions, thus facilitating the hiring of individuals that would otherwise be relatively less competitive candidates. 2 Beyond the initial question of determining how flexible workers differ from fully contracted workers in terms of professional qualifications and personal characteristics, I explore additional hypotheses with the intention of shedding light on the precarious employment situation of these workers. These are as follows: 2. Flexible workers and fully contracted workers differ in terms of how they view their employment, and how they are treated by their co workers and supervisors. 3. Flexible workers tend to receive lower pay and inferior benefits than fully contracted workers do. Hypotheses Regarding Gender In conjunction with the findings presented in Chapter 3 , the main hypothesis regarding gender is: 1. The shift towards more women in skilled positions at the Manaus Industrial District, is related to a growing gap between men and women in terms educational attainment, favoring the latter. Secondary hypotheses influenced by the literature on the feminization of factory labor and gende r in development are as follows: 2. When considering skilled and unskilled workers separately, women will have higher average levels of educational attainment than men. 3 2 . This assumption was naively based on the researchers own experience as a temporary worker in the United States. 3 . It is assumed that skilled workers will generally have higher levels of educational attainment than non skilled workers, regardless of gender. In the context of this s tudy, skilled workers refers not only to those individuals that perform leadership roles as supervisors and managers, but also machine
99 3. With regards to household finances and spending patterns, women workers with children wi ll be more likely to spend their income directly on their kids on such things as education, clothing, and healthcare than will men. Organization The rest of Chapter 4 is organized in the following manner. I first explain the methodology used in this study , including the terms used in the study, the sampling method, a description of the interview process, and the methods used in the analysis. Then I proceed to the findings, which is divided into two main sub sections. In the first, I discuss the findings re garding the differences between flexible and fully contracted workers. In the next, I relate the findings regarding gender. In the concluding section I synthesize the main findings of the study, acknowledging the limitations of the data collected, and sugg est future research related to the topic. Methods Terminology T he central concepts employed here in Chapter 4 are flexible worker and fully contracted worker. The former term has been translated from the original Portuguese, mÃ£o de obra temporÃ¡ria e terceirizada referring to individuals employed by a factory at days. Such tempora ry workers are only hired to meet production demands during peak times of the year; hence the temporary nature of their contracts, which facilitate their dismissal during subsequent lulls in production. Subcontracting is another common practice whereby fir operators, mechanics, technicians, and engineers. Educational attainment for the purposes of this study includes not only formal schooling, but also the number of training certificates individuals have received.
100 4 Although subcontracted workers generally perform the same tasks as their full time (and or temporary) coworkers, they remain employed through a third party for the duration of the l abor contract. Suframa includes statistics relative to temporary and subcontracted workers in its annual report, distinguishing them from fully contracted workers . The term fully contracted workers, is a translation from the Portuguese mÃ£o de obra efetiv a . 5 SUFRAMA de Desempenho do PÃ³lo Industrial de Manaus (2010), includes figures relevant to these different types of labor. In terms of factory workers, unskilled labor refers to individuals employed in quickly learned, labor inte nsive assembly tasks. In the production line, such jobs tend to be low paid and have high turnover rates. The term also refers to workers in the tertiary sector, who occupy positions that require relatively little certification and or training such as secu rity guards, receptionists, custodians, and food service personnel. In the context of this study these types of jobs are distinct from semi skilled and skilled labor. Not all flexible workers are necessarily unskilled . While the bulk of flexible workers c ontracted through tertiary human resource agencies are unskilled, there are 4 . Essentially recruitment agencies; in such cases the workers remain employees of the agency and not the factory where they work. Thus the factory avoids all the a dministrative costs that would normally accompany an increase in the employee roster. Such agencies play an important role in maintaining the labor order of the Manaus Industrial District by facilitating seasonal and short term employment as well as stream lining the recruitment and hiring processes. One of the key informants of this study is a partner in one of the largest of these recruitment firms in Manaus. Her insight into the nuances of recruitment and contract dynamics was an invaluable resource. 5 . The translation mÃ£o de obra efitiva to fully contracted workers could be deemed misleading considering that efetiva hour workweek in the US. However in the context of the MID it specifically refers to workers with standard or traditional contracts, Individuals directly employed by the firms thy work at, for an indefinite period of time.
101 various skilled technicians and specialists that are classified as flexible labor . This is because their employment within the factory is based on temporary contracts. 6 S killed wo rkers refers to those who perform tasks that require certification and training beyond what is generally taught in secondary school, and learned by entry level workers in their first few months of employment. In the context of this study this category incl udes machine operators, mechanics, quality controllers, supervisors, managers and administrative personnel. Such individuals usually acquired their skill sets and certification through either formal instruction (i.e. tertiary and certification programs) or through ascension within the hierarchy of the factory, typically over a period of months or years. These types of positions generally pay better then unskilled jobs, and are less often occupied by flexible workers . Sampling Method The research participa nts analyzed in this study were drawn through a snowball sample. Snowball sampling is a non probability sampling method whereby the These informants are then asked to refer other people who also meet eligibility criteria of the study. The sample expands rapidly as additional tiers of informants refer their respective friends, family members, and colleagues. It is a commonly used method by researchers within the discipli nes of anthropology and sociology who study less visible or otherwise less accessible populations (Atkinson, 2001) . While it can be an effective sampling method for individuals conducting qualitative research it is not without its risks. As researcher David Morgan (2008) points out, one of the potential risks of using the 6 . Two of the individuals I interviewed for this study are examples of semi skilled or skilled technicians that are also flexible workers.
102 snowball sampling method is that it can lead t (817). While there are a number of methods researchers can employ to assess the statistical accuracy of their sample, such a methods are limited in actual practice (Atkinson, 2001) . More often researchers just make do with the non random nature of the method. I deemed snowball sampling the most appropriate method for acquiring a research sample given my limited time and resources. Eligibility of Participants In order to qualify for the study, individuals needed to be working or had worked within the last year for a business located within the limits of Manaus Ind ustrial District, or that otherwise receives direct tax incentives from the Manaus Free Trade Zone. 7 The majority of the interviewees were employed as assembly workers in the electronics manufacturing pole, although a handful worked in the automotive or o ther sectors. Also, a number of participants worked in the tertiary sector, providing services to businesses in the industrial district. Seventeen of the assembly workers I interviewed had at one time or another been employed as flexible workers, whether t emporary or subcontracted. Among these about half were either currently employed as flexible workers , or had been so within the previous year. These individuals provided an 7 . While the vast majority of respondents worked for businesses that were physically located within the geographic limits of the Manaus Industrial District, in a couple of cases participants had worked for fÃ¡br icas de bairro (factories located in more residential areas). Although such manufacturers are not technically located within the industrial district, they still receive the tax incentives afforded by the free trade zone model, and their activities are coun ted alongside those of the firms located within the district in annual reports published by Suframa.
103 flexible workforce, so to speak. Locating Participants I found about a third of the respondents through a local NGO located in the center of the city that offers career placement and vocational training at no cost to job seekers. 8 Through this organization I found more than a dozen informants for the study, most of whom had recently lost or quit their jobs in the Manaus Industrial District. One advantage of interviewing unemployed individuals was their willingness to participate in the study. As one might expect, employed individuals were generally busier and far more difficult to schedule quality interviews with. Through networking and utilizing my pre existing relationships with local people, I was able to interview 28 individuals that were currently employed. Most of the respondents agreed to have their interviews recorded which allowed for careful transcription afterwards. Other than logging my research notes, I spent most evenings traveling around Manaus either on foot or by bu s in order to meet for interview appointments. With few exceptions most of the people I approached agreed to give interviews although at times, scheduling was difficult due to peoples' employment responsibilities. Because I arrived during the time of festa julinha , cultural festivals were frequent. On several occasions I was able to find additional informants by talking to people at social functions associated with seasonal festivities. I approached several companies located in the Manaus Industrial Distric t in the hopes of observing their factory operations and procuring additional respondents. 8 . The name of the NGO is O Centro de AutossuficiÃªnca SUD .
104 Although I was treated cordially, unfortunately, none of the four businesses I approached allowed me to visit. I also visited Suframa touri ng their campus and accessing their library. Unfortunately, they were unable to provide me with quantitative data on temporary and subcontracted workers disaggregated by industrial subsector, pay and gender. Since the sample is not random, I do not attemp t to generalize the descriptive statistics to the entire population of flexible and fully contracted worker s in the Manaus Industrial District. Nonetheless, the sample includes a valuable cross section of workers from various companies, positions and contr act types. Thus the data are useful for drawing preliminary conclusions about the differences between fully contracted and flexible workers, and the current gendered division of labor in the Manaus Industrial. District. Key Informants I also interviewed s ix individuals with extensive experience in hiring and managing workers in the industries of Manaus, whom I refer to as key informants. All six provided special insights on the nuances of employment in the Manaus Industrial District stemming from their vas t experience and relatively prestigious employment positions. All of them work in upper management roles and have participated in the recruitment and hiring processes of factory workers. Their responses and personal characteristics are maintained apart fro m the larger sample as they are not from the population of interest. They directly answered questions at the heart of the study, such as whether or not they considered there were any differences in terms of professional accomplishments, qualification, and job performance among flexible and fully contracted workers. Also, they answered specific questions with regards to the
105 proportion of females employed in skilled jobs and other nuances related to gender and employment in the Manaus Industrial District. Se mi Structured Interviews Most of the research data was collected through semi structured interviews although I also present data gathered through informal interviews and participant observation. All interviews were conducted during a five week period in Ma naus, Brazil, during the months of July and August 2013. In all, 45 interviews were conducted; 39 with factory workers, and six with key informants. I chose a semi structured format for the interviews because of time, logistical constraints, and the deduc tive nature of the research design (Bernard, 2000) . My questionnaire consisted of 63 questions. The first section of the questionnaire and l evel, marriage status, employment history, household income and expenditures, etc. The second portion consisted of more open ended questions such as: How would you describe your relationship with your supervisor? What do you like most about the company you work for? Would you prefer to work as a fully contracted worker rather than as a temporary/subcontracted employee and why? they view the Manaus Free Trade Zone as a developme nt project for their region. The personal characteristics I acquired for each respondent is particularly useful in considering how flexible workers differ from fully contracted workers in terms of demographic characteristics, work history and professional qualifications. The information acquired through more open ended questions lends itself to an overall
106 better understanding of how these workers view their employment and reveals how the cts negatively impacts a substantial group of workers. During the field research I also engaged in countless conversations with local people discussing the presence of the industrial district in Manaus, its effect on the local economy, the prominence of w omen in industry, and the local job market and professional culture. Although I do not discuss these conversations in detail, they do factor into the analyses of the phenomena in question. Coding Once transcribed, the interviews formed 191 pages of text, thus an essential step in my analysis was coding. I read through the pages many times to identify recurrent questions themselves served as preliminary categories to which I then assigned sub involves subdividing the data as well as 9 responses to this question could be placed within nine different response categories: pay, benefits, ambiance, relationships, technology and productive tasks, organization, freedom, opportunities for personal development, and unspecified. These response 9 . it generally
107 categories could then be further subdivided into micro c I created a matrix to record the frequency of response categories, and the stat istical software package SPSS for convenience. Using SPSS I was then able to quickly generate tables and other visual representations of the coded data. Sample Characteristics The research sample consisted of 39 individuals, ranging between the ages of 19 and 53 years of age ( Table 4 1 ). Twenty of the respondents were female and 19 were male. 10 Although key informants were excluded from the analysis it is worth noting that the majority of them were female. 11 Although I collected information on color (race) this characteristic ended up being less relevant to the study since there were no observable trends with regards to employment type or educational attainment between the participants when disaggregating by it. Following the prec edent established by IBGE, I offered survey participants the choice of any of the five color categories: brown, white, black, indigenous, and yellow. Eighty two percent of respondents identified themselves as brown or black, whereas only 7% self identified as white, and 5% as yellow ( Table 4 2 ). 10 . Gender is an important variable in my analysis since any information gleaned on the difference in perspectives between male and female workers regarding their experiences in industrial employment as well as their professional qualifications and labor histories is a useful compliment to the quantitative analysis regarding changing trends in the female composition of industrial employment discussed previously. For this reason I conscientiously attempted to keep the sample balanced between male and female respondents, across industrial sub sectors and employee types. 11 . Four of the key informants were female and two were male. Fem ale key informants provided especially rich data with regards to the gendered segregation of labor at the MID and generalized differences between male and female workers.
108 Fifteen of the 39 study participants (38%) are classified as skilled workers due to their specific occupation and or earnings ( Table 4 3 ). By disaggregating the data by skilled and non skilled forms of labor one can more clearly view the difference between flexible and fully contracted workers that are likely to occupy similar production roles. 12 Observed educational attainment of the sample is consistent with the bulk of literature on the labor force at the Manaus Industrial District. 13 With the exception of two outliers, all informants had completed secondary education ( Table 4 4 ). Brazilians regularly pursue certificate programs as a complement to their formal education in the school system. Examples of common certificate programs reported by respondents were basic and advanced computing, mathematics, electronic components handling, soldering, secretarial, and marketing. The major ity of respondents (87%) reported having completed at least one certificate program, with 32% reporting that they had completed three or more ( Table 4 5 ). With regards to higher education, 66% of respondents reported having at tended at least some college or vocational training after high school. Forty percent of those who had ever pursued some type of tertiary level education, concluded their degree. 14 and one with a graduate degree ( Table 4 6 ). 12 . physical product as part of their main job function, i.e. machine operators, technicians, and quality controllers. 13 . As far back as 1984 when DesprÃ©s conducted his 6 month study of Life and Social work at the graduates. In her 2001 study of female assembly workers in an electronics factory, Barbosa observed that 67% of t he workers she surveyed had graduated from high school, and all of those whom had not were concurrently enrolled as a condition of their employment (Barbosa, 2007: 95). 14 . Tertiary level degrees in this context are not limited to bachelors and graduate de grees but also include two year to 18 month trade school.
109 With regards to family, marriage and household size, only four respondents reported living in a household where they were the only adult. More than half (56%) reported living with a s pouse or partner and the remainder of the individuals reported living with parents, other adult relatives, or roommates ( Table 4 7 ). Nearly a third of respondents (28%) reported being married, and 56% reported having one or mor e children, although in several cases the children did not currently reside with them ( Table 4 8 ). The women in the research sample conform remarkably closely to the population of female workers in the electronics subsector in Manaus per the 2010 census data. As shown in Tables 4 9 and 4 10 , 54% of the women employed in the electronics subsector that year were between the ages of 19 and 30, and 77% were unmarried. 15 The respective percentages among the female portion of the sample are, 55% and 70%. Findings Flexible vs. Fully Contracted Workers Differences in education Much of the literature on factory workers in the Manaus Industrial District indicates that a high school education is usually a prerequisite for employment (Barbosa, 2007; DesprÃ©s, 1991) . A key informant, who works as a labor recru iter for factories in the district, indicated that all of the firms she works with require that production workers have a high school diploma regardless of the nature of their 15 . Refers to official marital status. Cohabitating individuals that are unmarried are counted among the single population in the relevant table.
110 contract. She elaborated that while in decades past some firms were more lenient with regards to this requirement, this had changed within the last 10 years. With the exception of two individuals, all research participants had graduated from high school. Although both of them were flexible workers constituting 25% of the eight flexib le workers interviewed one should be cautious about generalizing their educational situation to the entire population of flexible workers in the Manaus Industrial District. One of the individuals who had not graduated from high school was currently enrolle d, and planned to graduate within the next year. She explained that her employment status remained contingent on her enrollment since the company had a policy of only hiring high school graduates. Her official status was as a temporary worker, whereby she was employed directly by the factory under a ninety day contract. 16 She was hopeful that upon the conclusion of her contract she would be fully contracted . The other outlier was a 21 year old woman who had only completed grade school. Although not employed at the time of the interview she had previously worked for an informal temp agency where she performed basic assembly tasks for various electronics manufacturing companies in Manaus. Her work experience at the district was far from the norm considering tha t she was not employed with a work card ( carteira assinada ), was not unionized, and was compensated daily rather than bi monthly by her employer. Her situation as an informal subcontracted assembly worker was not unlike the subcontracting observed by Bener Ãa (1987) in her study of subcontracting in Mexico City. Although this case was an outlier in terms of employee benefits, compensation, 16 . Her case is included as that of a flexible worker .
111 and over all experience, I include it in the analysis since it provides a valuable glimpse into the phenomenon of inform al contracting. Overall, the sample conforms to both the literature, and opinion of key informants that suggest that employment in the factories of the Manaus Industrial District is generally contingent on having completed high school. There also appears to be little difference between flexible and fully contracted workers in terms of college enrollment and graduation. As displayed in Table 4 11 , 42% of fully contracted workers had never attended college compared to 50% of flex ible workers. In both groups 13% had graduated from a trade school whereas 13% of fully contracted workers claimed that they were working towards a trade school degree, compared to none of the flexible workers. Ten percent of fully contracted workers were college graduates whereas 19% reported that they had completed some years of college, but had not yet graduated. The most notable difference then is between the incidence of fully contracted workers and flexible workers who had graduated from college. As m ight be expected, workers in skilled positions are better represented among fully contracted rather than flexible workers, and skilled workers were frequently among the more educated research participants. 17 In order to better account for the differences i n educational attainment between flexible and fully contracted workers with regards to tertiary level education, I performed a second analysis excluding skilled workers. The results, shown in Table 4 12 , indicate that there is minimal difference between unskilled flexible and fully contracted workers, in this regard. There was little difference between the proportion of unskilled fully 17 . As displayed in Table 4 12, 80% of trade school graduates and 66% of college graduates in the sample can be classified as skilled workers
1 12 contracted workers (61%) and unskilled flexible workers (67%) that had never attended college. Although no flexible workers indicated that they were enrolled in a trade school program, nor that they had graduated from any tertiary program, 33% indicated that they had completed some college, compared to 28% of fully contracted workers who had likewi se attended but not completed. 18 Although one of the fully contracted unskilled workers had already graduated from college, he is best considered an outlier, since he himself confirmed that his situation was far from the norm. Although the sample of unskil led workers is fairly small, and one should be cautious about generalizing these findings to the entire population of unskilled workers in the Manaus Industrial District, these findings substantiate the comments of key informants, suggesting that there is negligible difference between flexible and fully contracted workers in terms of educational attainment. With regards to certificate programs, I did not initially find any stark difference between flexible and fully contracted worker s in terms of the numbe rs of certificate programs they had completed. 19 This was surprising as I expected fully contracted workers not only to have superior qualifications as a prerequisite of their employment, but to also have increased access to training programs funded by thei r employers. While many of the skilled fully contracted workers I interviewed indicated that they had acquired training certificates through their employment, this was not the case for any 18 . This figure of 28% is derived from combining the percentages of fully contracted workers that had been enrolle d in, but not concluded, college (11%) or a trade program (17%). 19 . One of the things I learned while about the professional environment in Manaus, is that individuals and employers generally place a lot of value on training programs that are administered through weeklong or sometimes month long workshops. Upon completion the participant receives a training certificate. Job candidates are counseled by employment agencies to list the trainings programs they have completed on their rÃ©sumÃ©s.
113 unskilled workers. This suggests that skilled employment may be cor related with a larger human capital investment on the part of firms. Once disaggregated by skill level, however, there was a pronounced difference between fully contracted and flexible workers with regards to the number of training certificates they had acquired. As shown in Table 4 13 , 25% of fully contracted worker s indicated having completed 3 or more certificates, the same percentage as flexible workers. However, once skilled workers are excluded ( Table 4 14 ), the percentage of flexible workers who had earned three or more certificates increased to 33%, compared to 16.7% of fully contracted workers. Additionally 50% of flexible workers had earned two certificates, compared to only 25% of fully c ontracted workers. Worth noting, all unskilled flexible workers, included in the analysis, had earned at least one certificate, whereas a third of fully contracted unskilled workers had not. These findings are consistent with the opinions of key informants who indicated that flexible workers frequently had superior rÃ©sumÃ©s to their fully contracted counterparts. One key informant attributed this to the fact that most subcontracted and temporary employees had experienced unemployment in the recent past, and had time to seek out additional training and qualification while out of work. Based on the analysis of characteristics of the research sample and qualitative data gathered through interviews, there seems to be negligible difference between flexible and ful ly contracted production workers in terms of their education and professional qualifications. These findings contradict the original hypothesis, which was based on the premise that flexible workers tend to be less educated and less experienced than fully c ontracted workers .
114 Differences in pay and benefits The main differences with regards to flexible workers and fully contracted workers are manifest in the pay and benefits they receive. All workers, regardless of contract type, reported receiving the benef its of transportation and meals during their workday. Fewer flexible workers reported receiving health or dental care from their employers. As shown in Table 4 15 , fewer flexible workers benefited from end of year bonuses and p harmacy discounts. Of the 18 individuals interviewed who had ever worked as subcontracted employees, only three reported receiving health plans from their employer during the duration of their subcontracted status. Of the nine participants that had exper ienced employment as temporary workers, four of them reported receiving healthcare during their temporary contract. Although the data suggests that flexible workers historically have not received healthcare from their employers, there is evidence that thi s trend is changing. According to two of the key informants , healthcare has become increasingly mandatory even for temporary and subcontracted workers. They attribute this to progressive demands of O Sindicato de MetalÃºrg icos do Amazonas ), and relatively new federal labor legislation (Key informant, personal communication, July 22, 2013; Key informant, personal communication, July 26, 2013). One of the informants indicated that approximately 40 percent of the subcontracted employees hired by her recruitment firm currently receive healthcare benefits (Key informant, personal communication, July 26, 2013). It is worth noting, however, that such health plans are reportedly of inferior quality to the plans provided to their fully contracted em ployees. Also worth noting is that the few subcontracted and temporary employees who reported
115 receiving health plans from their employers had all been employed as such within the last year. The relatively recent nature of their employment supports the noti on that this is an emerging trend. Informants from both employment categories frequently claimed that flexible workers receive inferior wages to their fully contracted counterparts. As Suframa does not provide wage data disaggregated by employee contract n or manufacturing subsector we are unable to confirm this from the available workforce statistics. Furthermore, because the research sample does not include any cases of flexible and fully contracted workers who concurrently worked for the same firm in the same job function, we are unable to empirically examine whether or not this is true. Nonetheless, when disaggregating the sample data by employee type, the mean monthly earnings of flexible workers are indeed inferior to the earnings of fully contracted wo rkers, with fully contracted workers earning R$ 1259.90 and flexible workers earning R$853.71 on average, respectively ( Table 4 16 ). Because skilled workers are better represented among the fully contracted workers one would ex pect this difference to be pronounced. After excluding the cases of skilled workers from the analysis, however, the difference amounts to about R$45 or only 4% ( Table 4 17 ). The difference in benefits received by flexible and f ully contracted workers was one of the first things mentioned by respondents when I inquired if they perceived any difference in the treatment by their employers. The vast majority of informants stated that temporary and subcontracted workers received inf erior benefits compared to their fully contracted counterparts. All of the informants stated that given the option, they would choose to work as a fully contracted rather than as a temporary or subcontracted
116 employee. The reason they most often gave for th 2013). He was referring to the relative lack of j ob security he felt as a flexible worker. In the majority of cases, workers hired under temporary or subcontracts are only employed as such for a matter of weeks or months in order to meet seasonal spikes in production. Their contracts are set to expire wi th the projected seasonal decline in production. Many research participants, who had previously worked as flexible workers, indicated that they experienced this process first hand. fully positions. While this may be true to a certain extent, various research participants criticized employment in the district in general terms, claiming that the mo del provided them with little job and financial security, and that jobs were generally of low quality. These themes emerged when participants were asked to share their opinion of the Manaus Free Trade Zone as a model for regional development (See Table 4 18 ). Fifteen percent of research participants generalized that jobs at the Manaus Industrial District lack security, including 13% of fully contracted workers and 25% of flexible workers. Additionally, 21% of research participa nts described the jobs generated by the MFZ model as generally being of low quality, 19% of fully contracted workers and 25% of flexible workers. Among the participants that criticized the MFZ, claiming a general lack of job security, was a former fully c ontracted worker that had lost his job at a motorcycle plant
117 6 months earlier. He shared details about his own declining health, which he blamed on the work routine and ultimately on his supervisor. He explained the circumstances that led to his dismissal, wherein he contracted an infection at work, and was given time off to recuperate. Upon his return to the factory, he was promptly laid off. Although his employer claimed that the lay off was impartial, he is convinced that the company targeted him because of his health condition (Assembly worker, personal communication, July 2, 2013). He suggested that employers in the Manaus Industrial District frequently layoff individuals with health problems despite legislation forbidding the practice. Twenty five perc ent of currently unemployed research participants indicated they had lost their previous employment due to a lay off. These responses are displayed in Table 4 19 , disaggregated by employee type and gender. Forty four percent of the unemployed individuals indicated that they quit their last job at the industrial district on their own accord. Among the reasons participants most frequently gave for having done so, were being tired of the physical routine of factory employment, the desire to pursue other employment alternatives, and a need to focus on their education. Only one former flexible worker gave this latter reason for why she had left her employment, to focus on her studies (Assembly worker, personal communication, July 15, 2013). An analysis of individual characteristics and qualitative interview data suggests that substantial difference exist between flexible and fully contracted workers in terms of the benefits they receive from their employers. These findings confirm t he hypothesis stated at the beginning of Chapter 4 . Also differences appear to exist between fully contracted and flexible jobs in terms of perceived stability. Various research participants
118 indicated that flexible positions provide less security than full y contracted positions. of research participants characterized fully contracted and flexible positions alike as unstable and insecure. Differences in treatment Another important question with regards to flexible employment is whether or not flexible workers experience discrimination in their jobs beyond inferior benefits and pay. I asked respondents whether they perceived different treatment on the assembly line. Most respondents (90%) mentioned that they did not perceive unequal treatment on behalf of supervisors and management; however, 80% mentioned that they noticed discrimination on the part of fully contracted employees (See Table 4 20 ). I interviewed one flexible worker who had once been fully contracted by the firm where he had worked during the preceding 90 days under a subcontract. He shared an experience in ck up the pace buddy, pick up the pace, you're still only a 'peon' you'll need to work a lot harder if you view this criticism from the fully contracted worker was not o nly unfounded, it was inappropriate since the fully contracted worker technically carried out the exact same job function and had no leadership role or authority over him. The idea that fully contracted employees might feel threatened by flexible workers i s easy to understand when one considers that many temporary and subcontracts are given to the workers as initial trial periods after which the flexible employee has the possibility of being fully contracted . Confirming this, a production line manager ment ioned that she frequently hired flexible workers to replace less productive
119 fully contracted employees once the initial contract period expired (Manger, personal communication, July 21, 2013). She suggested her company uses subcontracted labor not only to increase production during peak times of demand, but as a means of testing job candidates before bringing them on full time. Other informants echoed her sentiment that fully contracted workers justifiably view flexible workers as competition (Key informant , personal communication, July 11, 2013; Key informant, personal communication, July 26, 2013). Other themes concerning the difference between flexible and fully contracted employees that regularly surfaced in the interviews was the idea that flexible work ers are frequently blamed for mistakes in the production line whether or not they are actually responsible. According to several respondents, flexible workers are frequently blamed for problems caused by fully contracted employees. Respondents also cited institutional discrimination as an observable difference between the treatment of flexible and fully contracted workers. Institutional discrimination in the context of this study ( Table 4 20 ) is a term referring to a variety of company policies that result in differential treatment of different classes of employees. mobility on campus (not permitted to move about as freely as fully contract ed individuals), use of different uniforms, and segregation during lunch hour whereby flexible workers are not allowed to eat at the same time as fully contracted employees. Although respondents cited examples of institutional discrimination less frequentl y, those that they did mention suggest the nuanced treatment received by employees in the industrial district.
120 Professional goals I found little difference between flexible and fully contracted employees in terms of their future goals. Most respondents (7 4%), regardless of their employment status indicated that they planned on continuing their education. Thirty eight percent of respondents indicated that their future plans included working in another profession outside the industrial district, where as 41% indicated that they would like to continue in or return to the industrial district given the chance. Four individuals expressed that they only planned on continuing in their current employment at the district for as long as they were able to and expressed no desire to move up within the company (See Table 4 21 ). The cases where responses between flexible and fully contracted differed most were continue employment in the industrial district, compared to 45% of fully contracted workers. This may be because fully contracted workers (many of whom occupy skilled jobs in the sample) enjoy a relatively better form of employment in terms of security and benefits. When one considers the universal preference of research respondents for fully contracted over flexible employment, it is unsurprising that no flexible workers in dicated that they planned on maintaining their current employment status. This is especially fitting considering the usually short contracts. Gender Analysis I approached this study with a number of questions related t o the influence of gender on employment at the Manaus Industrial District, the most important being whether female participation in skilled positions increased in recent years. Although
121 implies having taken place within the employment history of the individual being interviewed. Thus if a participant indicated that they had witnessed a sustained shift towards more women in skilled and leadership capacities, that constitutes as having happ district, and those that did, who indicated having seen a shift towards the employment of more women in skilled jobs, indicated that the shift has been within the previous 10 years. Also considered is the information participants supplied regarding the gender of their overseers (leader, supervisor and or manager) and the observed proportions of men and women working alongside respondents in their particu lar work areas. 20 The insights of key informants were especially important to this section of the analysis because these individuals on average had more than 10 years of experience working for and with businesses in the district, and tended to exemplify mor e holistic views. Other important gender issues are related to demographic and financial differences between factory respective education levels, living arrangements, employment categories, and spending plans disaggregated by gender. 20 . In Brazilian in dustrial employment, the term work area (translated from the Portug u ese Ã¡rea ), refers to productive tasks the worker performs generally within in certain place in the factory at specific stage in the production process. As a product is manufactured in the factory it generally passes through various physical areas each associated with a different stage in the production process. For example in the production of cellphones, component assemblers, and quality controllers occupy different stages in the producti on process and figuratively different work areas.
122 Educational attainment Contrary to the literature that indicates that female educational attainment levels in Brazil have generally sur passed those of males, the women workers, interviewed in this study, were on average less educated than the men. The reason the sample data does not correspond to the literature, nor the census data analyzed previously, may be due to sampling bias. As show n in Table 4 22 , 90% of female respondents indicated that they had graduated from high school compared to 100% of male respondents. With regards to college enrollment and education, 50% of female participants indicated that they had never attended college, including trade schools whereas only 37% of male participants indicated the same thing. One fourth of female respondents indicated that they had previously attended a post secondary program in a trade school, whereas 21% o f men indicated similarly. Twenty five percent of women had at least some college compared to 37% of men. Only one of the participants, a male, indicated he had concluded a graduate program, the only person in the sample that reported having some graduate training. 21 Professional qualifications When it comes to professional qualifications, as operationalized by number of training certificates earned, the difference between male and female achievement is more nuanced. I quickly learned at the employment cen ter where I volunteered part time during the duration of the study, that training certificates play an important role in job candidate evaluation in Manaus. Various institutions, both state run and private, 21 . All six key informants, four of them female, had college degrees, one of them had an advanced graduate degree, and another one was working on a graduate degree.
123 provide educational training programs to students and jobseekers. Most of these programs last from one month to 90 days and upon completion the participant receives a certificate that s/he is counseled to keep in the case of future demand for verification from an employer. 22 Thirty one percent of the fem ale respondents reported earning 3 or more training certificates, similar to male participants, 33% ( Table 4 23 ). 23 More than half the women 54% indicated having earned two certificates, compared to only 17% of the men. No wome n indicated having received only one certificate and 15% indicated that they had not received a certificate, compared to 39% and 11% for men, respectively. Women in skilled positions I found some evidence regarding the growth of female participation in sk illed of their respective supervisors and the percentage of females employed in high status positions in their companies, suggests that this is the case. Also, several key informants indicated that they had witnessed growth in the proportion of females in skilled positions in recent years. women employed in their respective work areas suggest that occupational segregation 22 . There are myriad programs and subjects in which one can earn a training certificate. Among those most frequently reported by participants were, certificates in basic computer literacy, soldering, telephone switchboard operation, and handling of electronic components. There is so much variation however among the types of courses participants reported, that it would be impractical to display this information in a table. 23 . Originally the researcher did not intend to report information on number of certific ations respondents completed, for this reason during the data collection process he failed to collect this information for eight of the respondents unfortunately seven of which were female. Nonetheless during the analysis it was determined that this data a lthough only available for part of the sample is valuable to understanding differences in male and female educational professional attainment.
124 remains prevalent in the Manaus Industrial District, a more careful analysis suggests that more females are finding employment in occupations traditionally dominated by men. When asked whether there were more male or female worker s in their respective work areas, sixty five percent of women responded that there were more females and 63% of male respondents indicated that there were more males ( Table 4 24 ), which is suggestive of occupational segregation by gender (Chant, 2003) . Perhaps more interesting, however, is that 35% of female respondents indicated there were more males than females in their respective work area. Many of these females occupied skilled positions either as machine operators, supervi sors or quality controllers (See Table 4 3 ). While this indicates occupational segregation, the fact that more than a third of female respondents (35%) are employed in capacities traditionally occupied by males, suggests that w omen in such positions in the district are not a rarity (Spindel 1987) . to coincide with their own gender. As Table 4 25 shows, 55% of females indi cated that their overseer was also female whereas 54% of males indicated that their overseer was male. Forty percent of females reported that they worked directly under a male overseer and 38% of males indicated they worked under females. These figures su ggest that gendered segregation by manufacturing subsector extends to leadership positions. 24 This trend is further substantiated by one of the key informants who worked as a production manager for an appliance manufacturer. He indicated that in his 24 . In a manufacturing subsector dominated by one gender, supervisors are generally going to be selected from the dominant gender, i.e; in the Honda motor plant where the vast majority of employees are male, so are most of supervisors.
125 company female supervisors oversaw manual assembly processes, and male supervisors oversaw more mechanized stages of production. He explained that having a female supervisor over manual assembly was practical because that was an area of production dominated by fe males anyway (Key informant, personal communication, July 11, 2013). In spite of this general trend, a large minority of the males interviewed (38%) actually had females for supervisors. While this can partially be explained by the fact that some of these males worked in a production areas within a subsector traditionally dominated by females, this was not the case for all of them. 25 One of the key informants, who works as an operations manager for a multi national manufacturer of electronics, indicated tha t she predominantly promotes female workers to leadership roles within her jurisdiction. She stated that she does so because ger, personal communication, Aug 13, 2013). Another research participant who worked at a toy factory indicated that male believed the transitive nature of most men was the reason females were more readily worker, personal communication, July 15, 2013). 25 . Two of the male informants interviewed worked in production areas that had a majority of male workers however, their supervisors we re female (Assembly worker, personal communication, July 4, 2013; Assembly worker, personal communication, July 31, 2013 ) .
126 A key informant who recruits workers for the electronics subsector, indicated that she prefer red hiring females for the assembly line, because she had learned, during her among the first to quit their jobs or be excessively tardy and absent in their posts (Labor recruiter, personal communication, July 26, 2013). Similarly an HR manager for a multinational manufacturer of cellphones noted a ung men of Manaus. She commented that they are generally less likely to pursue education and job training than their female counterparts and are less likely to dedicate themselves to fixed employment (Human resource manager, personal communication, July 22 , 2013). She also provided valuable insights into trends regarding the feminization of skilled positions at her own company. While she conceded that female workers are still shift to of only male mechanics, today 40% of the mechanics are female, and in many cases commun ication, July 22, 2013).. Gendered division of assembly processes When asked about the proportion of females to males in the companies they worked for, various research participants referred to a traditional division of production
127 tasks by gender. 26 Also several key informants indicated that manual assembly processes ( inserÃ§Ã£o manual ) tended to be dominated by female workers . When asked why they thought females were concentrated in manual assembly jobs, most participants cited stereotypes of females having smaller hands and greater dexterity. One male participant (who worked in female dominated assembly line) went average for a man, which is why I am so well suited worker, personal communication, Aug 14, 2013). As shown in Table 4 26 , 13% of all research participants referenced females as generally being better suited for assembly work because of their p 27 Also references to male stereotypes of being stronger and thus better suited to assemble heavy appliances and motorcycles, were frequent. Judging from these interviews, notions of male superior strength and femal e superior dexterity are deeply engrained in the minds of those who work in manufacturing in Manaus. Like many stereotypes, there is probably some truth to these notions yet by holding to them individuals and companies are likely to end perpetuating unnece ssary discrimination. Considering the prevalence of stereotypes regarding the suitability of workers for certain tasks based on their gender, it is not surprising that women who successful defy 26 . Table 4 26 shows gender related themes that commonly emerged throughout the interviews. Twenty three percent of respondents ment ioned that women tended to be concentrated in manually intensive production tasks within the electronics sector. Most of these respondents mentioned this when asked what was the proportion of women to men in the company they worked for. 27 . The converse o f this stereotype was also expressed by a number of research participants. Seven percent of % of the respondents indicated that men are not generally fit for manual assembly work 26).
128 such stereotypes take great pride in their accomplishment. Th ree of the women I interviewed expressed pride at working in a professional category that was generally dominated by men. One of the key informants, a manager within an electronics ment... part of me wants to continue progressing and see if I can't become general manager of recently worked as a quality control specialist for a manufacturer of motor cycle communication, Aug 13, 2013). Her friend who worked for the same company, only in a female s true she is the only woman, and the one of the respondent who was about to start a new position at Honda Motors on the production li (Auxiliary services worker, personal communication, July 25, 2013). Sexual harassme nt and exploitation Although I did not question any participants on the prevalence of sexual harassment at their jobs, a few respondents shared relevant anecdotes, without being prompted to do so. Such comments emerged organically in the interview process. One of the young women who had worked as a quality controller in a male male coworkers began to pass around a compromising photo of [her] they had stolen without [her] p
129 department who gave her the option of leaving the company with severance pay and unemployment benefits. Ultimately, she opted not to denounce her coworkers and instead chose to quit he r job, taking the severance pay and unemployment benefits offered her. Unfortunately five months later she was still unemployed and searching for more tragic about her case i s the fact that she was one of the female informants who expressed pride for working in a male dominated environment. Furthermore, she was employed in a skilled position, one that paid better than for entry level assembly work. When starting a new job, she is unlikely to begin at the same pay grade. Another female (age 26), working in lower management for a plastics injection company, decided to quit her job because of an indiscretion of her male supervisor. Although he never objectified her personally, sh e witnessed him viewing pornography on a company computer during the workday. She is an expressly conservative her interactions with him. When she went to speak to her HR department about his use of company property, and the level of discomfort it caused her, they provided her with the option of amicably ending her employment with severance pay and unemployment benefits rather than pursuing the issue further. Her willingnes s to walk away was likely influenced by her confidence that she would be able to find other employment based on her strong credentials, and the relative affluence of her parents. One of the male informants who considers himself an erudite on social issues, discussed the prevalence of sexual and romantic liaisons between coworkers at the industrial district. He claimed that while relationships between male supervisors and
130 their female employees were commonplace, this was only one aspect of a larger sexually free subculture. While taking care to distance himself from any romantic employees (Ass embly worker, personal communication, August 14, 2013). Implicit in his assertion, is the idea that the line between sexual exploitation and consensual relationships is frequently blurred. Future plans by gender When asked about their future plans, women were more likely to mention educational goals. As shown in Table 4 27 , 85% of respondents indicated that they wanted to either finish a degree or pursue additional education. Although the majority of men also mentioned educati onal goals, a smaller percentage (65%) did so. While these findings are consistent with the literature on the growing educational gap between males and females in Brazil, it is worth mentioning that the prevalence of this response among both male and femal e participants may be partly due to sampling bias as more than a third of research participants were contacted through an employment/job training office. 28 A larger percentage of male than female research participants mentioned that their future professiona l plans included staying in their current employment at the district. Forty five percent of men indicated that they planned on continuing in industrial employment, and 21% indicated more specifically that they planned on maintaining their current post inde finitely or until retirement. Only 35% of women indicated that they 28 . Individuals w ho value education may be more likely to frequent and use the educational services offered by the LDS Employment Resource Center.
131 planned on staying in industrial employment and none of the women interviewed indicated that they planned on staying indefinitely in their current employment at the district. Forty five p ercent of the females interviewed actually indicated that their future employment plans were to leave industrial employment. Only 31% of males indicated similarly. Among the future career paths female respondents most commonly cited as wanting to pursue we re careers in education, healthcare, and administration. How Workers View the Manaus Free Trade Zone Manaus thirty years ago he found that many assembly workers viewed their empl oyment as inferior to positions in the public and commercial sectors. Also, many individuals saw their jobs in the district as a transitory rather than a long term career path. When I asked respondents how they viewed the Manaus Free Trade Zone as a model for development and what their future plans for employment were, many of the attitudes expressed were similar to those recorded by DesprÃ©s. For example, one give you s ecurity... The only way to be someone is to study and qualify for something As I listened to workers express their thoughts on the MFZ model I recognized some consistent themes. I coded for eleven themes/categories that represent the variety of ways workers view the industrial district. These themes are displayed in Table 4 18 , the most prevalent of which is that the MFZ is essential because of the jobs it provides to the residents of Manaus. The next most common theme to emerge was that the types of jobs available in the industrial district are generally of low quality, with 21%
132 of respondents mentioning this. A few other interesting themes expressed by respondents have to do with the MFZ as a force for the preser vation of the rainforest (3%), the need for businesses to invest more in technology and innovation (11%), and a scarcity of skilled labor in Manaus to meet the demand of higher quality jobs in the district, (5%). Several respondents used the Portuguese eq to describe their relationship to their employer. When asked how she viewed the MFZ, one of the key informants expressed that although the industrial district provided her with numerous opportunities for personal growth and an acceptable way to earn a living, was required to spend so many hours away from home and her three children (Operation manager, personal communication, August 13, 2013). This key informant's sentiments are interesting because she was relatively successful in her career in the industrial sector, having completed a graduate degree in production management, earning various certifications, and currently earning a salar y that placed her in the top 8 percent of earners in the sector. One can only wonder if her feelings of regret stem from deeply engrained societal expectations regarding the role of women as primarily responsible for reproductive labor. This seemed to be t he case for some of the female union activists Iraildes Caldes (2005) interviewed in her study of the historical construction of female identity in Amazonas. Caldes's study identified one participant in particular, that had dedicated two decades of her lif e to actively leading the Amazonas Metal Workers Union. Somewhat disillusioned with the experience,
133 The Household Myriad studies exami level (World Bank, 2010). Policy makers design programs that take into consideration providers for children, but because of the literature suggesting that women tend to spend a larger portion of their income on their children then do men (Duflo, 2011) . it often reduces her autonomy (Dalal, 2011) . This is especially true in situations where her partner is abusive, negligent or expropriates her earnings. Thus living arrangements along with income and asset ownership have important impli empowerment. In this section I examine differences between male and female factory workers in terms of living arrangements, earnings, and how they spend their income. While a principle focus of this research is to elucidate differences between flexible and fully contracted workers in terms of qualifications and perceptions, another objective is to identify differences between male and female factory workers. As few studies of the Manaus Free Trade Zone directly compare interview data wit h male and female workers, I seek to make a contribution to the existing literature and gain insight into development at the household level. Marital status, living arrangements F irst we consider differences in the marital status and the living arrangements of the women and men interviewed. The majority of respondents (70% of women and 63%
134 of men) were single. 29 A larger percentage of males (37%) than females (20%) were married. Bot h of the respondents who reported being divorced were women (10% of female respondents) (See Table 4 28 ). Twenty one percent of single females indicated that they lived alone, compared to 14% who indicated they cohabitated with a partner. In contrast only 5% of single males lived alone whereas 42% lived with a partner (See Table 4 29 ). While these findings are likely due to sampling bias it is also possible that employed females are less likely to ac cept cohabitation than marriage, which may imply a more lax commitment on the part of their spouse. Additionally, the fact that more females are single and non cohabitating than males may suggest that women prefer the autonomy that living without a spouse or partner might provide them. As displayed in Table 4 29 , 20% of females reported that they were the only adult in their household at the time of their employment, compared to only 5% of male participants. 30 Conversely 63% of m ales indicated that they lived with an adult partner or spouse, while only 30% of women indicated the same. Forty five percent of female participants indicated that they lived with their parents or other adult relatives (generally siblings or grandparents) while 31% of males reported similarly. One of the female participants indicated that she lived with a non kin roommate, whereas none of the male informants did so. Generally these findings are consistent with the literature on female factory workers in Ma 29 . had once been marri ed but at the time of the study were divorced, indicated their marital status as divorced. There were no widows or widowers included in the study. 30 . int erview. Only one of the respondents indicated that her living arrangement had changed in the months after she lost her job. During her time employed she lived alone, thereafter she began living with a friend.
135 (70%) were single, both in a domestic and legal sense. I was surprised to find a relatively larger percentage of women, 20% living alone as compared to men. 31 The fact that a larger portion of female participants than male participants live with their relatives (45% and 30% respectively) may be partly explained by their age distribution. As displayed in Table 4 1 , similar proportions of males and females fall in the youngest age category (19 25 years); smaller proportions of males fall within the next two age categories (26 40 years). More than a quarter of males are in the older age category (41 years or older) compared to only 5 percent of females. Worth noting is the one female participant in the oldest age group was divorced and lived alone whereas four of the five men in the oldest age group lived with their spouse or partner. Budget and finances Substantial differences in household budget expenditures are found when disaggregating by both living arrangement and gender. As in Table 4 30 , expenditures on rent/mortgage were highly dependent on household composition. Seve nty five percent of women who were the only adult in the home, paid their own housing costs. paternal grandparents. Also worth noting, the one man who reported being the o nly adult in his home, indicated that his sister generously paid his housing costs. The category N/A in the case of housing refers to cases where individuals lived in a fully owned home, where neither they nor any of their cohabitants regularly paid rent o r mortgage. 31 . Although these findings cannot be generaliz ed to the entire population of workers, the greater incidence of females living alone may indicate a conscious choice on the part of female workers for autonomy, or conversely the fact that they live alone may necessitate their employment in the first plac e.
136 Only 30% of partnered women reported living in a fully owned home, compared to 58% of male respondents. There were two cases where individuals living with their extended family were the primary people responsible for paying rent, both of them were females employed in skilled positions. 32 None of the males that lived with extended family reported paying rent or mortgage. In all six cases they lived in fully owned homes. The literature consistently demonstrates that women tend to spend a greater proportion of their income on their children than do men. I suspected this would be case when interviewing factory workers for my research. 33 In order to assess spending patterns I questioned each informant on their household expenses, asking who in the ho usehold was the primary person responsible for paying each bill, including those associated with the children, such clothing, medicine, school supplies, and toys. While I did not gather specific information on the cost of each bill (including monthly expen ditures on children), I did ask participants how much of their earnings remained for their personal discretion after they had paid all the aforementioned expenses. By dividing this amount by his/her total monthly earnings one can assess what percentage of related to their children. Although it would have been more effective to gather information on specific monthly expenditures on children and divide those by total monthly earn ings, these data provide a general idea how mothers and fathers spend their earnings differently. 32 . In both of these cases the respondents also indicated they highest salary in the home. 33 . Because food, medical, and education expenses are already each considered separate categories, ccessories was included in the interview questions related to household expenditures.
137 As shown in Table 4 31 , the majority of participants, 69%, including both genders, indicated that they had living children and mo re than half of these (15 individuals), reported that they lived with minors in the home ( Table 4 32 ). 34 To assess differences in spending patterns by disaggregating by gender, I limited the analysis to only respondents who actu ally live with their minor children. 35 All households in the study that were headed by a single parent were headed by women. As shown in Table 4 32 , three women indicated that they were the only adult in the household with und erage children. Four of the five men interviewed who lived with their children were either married or co habited with the mother of their child(ren)]. One single father lived with his child, however, he resided with his parents, and his mother played an im On average women spent a higher percentage of their monthly earnings on household expenses than did men. As shown in Table 4.33 left for their personal use, after pay ing bills associated with their household and family, was only 14% of their total earnings, compared to 23% for men. After excluding skilled workers from the analysis, the difference is even more pronounced, with unskilled female workers spending on averag e 91% of their income on household expenses, compared to an average of 77% for unskilled male workers. It is worth noting, however, that there is a small wage gap between male and female working parents, with average monthly earnings of surveyed mothers at 96% the earnings of fathers ( Table 4 34 ). This difference is more pronounced for unskilled workers, however, with mothers earning 34 . Minor: 17 years old or younger. 35 . It is expected that single parents, whether or not they live with extended family are generally more likely to report themselves as the primary spender on clothing and other items for their children.
138 only 86% of what fathers earn ( Table 4 35 ). Considering the wage gap between male and female workers with children and the small sample size used in the analysis, one should be cautious about applying these findings to the entire population of workers. Nonetheless, these figures suggest that women tend to spend a larger proportion of their income on their household expenses (including their children), as supported by the literature (Duflo, 2011). Severance Pay Consistent with the academic literature that addresses differences in spending habits between men and women for their children, are the findings regarding how participants spend their severance pay. Twenty two of the participants (about half) indicated that they at one time received severance pay in conjunction with losing a previous job ( Table 4 36 ). The actual amounts received varied substantially as did what they spent it on. The most common expenditure for participants regardless of gender (38% of females and 38% of males) was on liquidating previous debts, and paying for regular hous ehold expenditures including groceries, housing and utility bills. Fifteen percent of females reported depositing their severance in a savings account while no male participants indicated the same. Similarly 15% of females immediately spoke of their childr en when asked how they spent their severance pay, citing the purchase of referred to their children in the context of spending severance pay, however, 13% indicated th at they spent their pay on expenses related to their education compared to only 8% of females. Lastly, no females from the sample mentioned investing the money
139 in a business venture, whereas 14% of males indicated that they had done so. 36 These last finding wherein she found that men frequently invested the money they received via severance pay and unemployment benefits towards starting a home business. She did not indicate how females typicall y spent their severance pay (Scherer, 2004). Conclusion tax burdened societies, the Manaus Free Trade Zone enjoys a singular existence (Alper, 2012). It has been championed by successive presidential administrations as an answer to regional inequalities and rainforest preservation, yet its fundamental role as an employer surpasses its other functions. Unfortunately, it is not immune to the restructuring of global capitalism, wh ich is why there have been dramatic shifts in the size of the workforce and the adoption of more flexible forms of labor since the 1990s. Flexible labor at the MFZ best manifests itself today in the prevalence of subcontracted and temporary positions at t he factories in its industrial district. Despite workforce (between 7 and 12%), subcontracted and temporary workers are rarely addressed in the literature on the Man aus Free Trade Zone. My research sought to shed light on this relatively marginalized population, as well as explore issues related to occupational segregation by gender in the Manaus Industrial District. 36 . Although not included in the analysis concerning severance pay, one of the key informants indicated that she was in the process of using severance pay funds from her previous employme nt to invest in a business adventure. One of the male key informants also indicated that he had invested funds from severance pay in his side business.
140 My primary hypothesis regarding flexible labor wa s that notable differences exist between flexible and fully contracted workers in terms of their professional qualifications and demographic characteristics. Subsequent hypotheses were that, differences also exist between flexible and fully contracted work ers in terms of how they view their employment, and the treatment and benefits they receive. 37 The analysis reveals no negligible difference between flexible production workers and fully contracted employees in terms of their respective levels of education al attainment and professional qualifications. My data does reveal pronounced differences between flexible and fully contracted workers in terms of the benefits they receive from their employers, particularly regarding healthcare. Most research participan ts cited inferior benefits as the main difference between flexible and fully contracted employment and the primary reason they would prefer the latter. Other important elements of discrimination however are manifest in the treatment flexible workers receiv e on the assembly line from their fully contracted counterparts. Most participants viewed the use of flexible labor in the Manaus Industrial District mporary jobs are great for the businesses of 'the district' but bad (Machine operator, personal communication, July 13, 2013). A key informant who works as the HR manager fo r a prominent cellphone manufacturer similarly expressed, 37 . Treatment refers to how workers are treated by their overseers and coworkers.
141 22, 2013). Due to the demands of loc al labor unions, I predict a move away from subcontracted labor in the assembly lines of the Manaus Industrial District during the coming years. The Amazonas Metal Workers Union now requires that flexible workers be provided with health benefits and receiv e salaries equivalent to fully contracted workers. Once health plans and higher salaries are factored into the expenditures of companies that use subcontracted assembly labor, the costs associated with hiring through a recruitment or talent agency may not justify the continuation of using subcontracted workers. Further evidence of this is found in the interviews with several respondents, who indicated that they had seen a decrease in the presence of subcontracted employees in their factories' assembly lines . Additionally, a careful review of data published by Suframa on the percentage of flexible workers employed across the 21 subsectors of the Manaus industrial district, suggests a slight decrease in the proportion of flexible workers since 2010 ( Table 3 1 ). 38 My qualitative findings substantiate the hypothesis that there is a growing trend towards more females in skilled positions at firms in the Manaus Industrial District. Several informants indicated that they had seen the gro wth of females in skilled positions within their respective jurisdictions. Although the traditional division of industrial labor by gender persists in the Manaus Industrial District, with females overrepresented among the manual productive tasks in the ele ctronics sector, various female research participants worked in skilled positions and in departments traditionally 38 . As shown in Table 3 1 since 2011 the monthly percentage of flexible workers at the Manaus Industrial District has remained at below 10%.
142 dominated by men. The relative ease with which I was able to find these participants is evidence that they are becoming commonplace. My sec ondary hypotheses regarding gender that female workers average higher levels of educational attainment than male workers, was not substantiated by the qualitative analysis. Findings regarding differences between male and female workers in terms of average educational attainment were inconclusive. Even after disaggregating by skilled and unskilled workers the sample data showed more college and trade school graduates among male workers. These findings are inconsistent with the bulk of quantitative research on the subject Also inconclusive were the findings with regards to household expenditures. While I did not collect sufficient data to determine whether or not male or female workers spend a greater portion of their earnings specifically on their children, female respondents tended to spend a larger share of their earnings on household expenses than male respondents. Additionally, women were more likely to indicate that they spent their severance pay on their children. These findings are reminiscent of the literature indicating that females tend to spend a greater proportion of their income on their children then do men. Other differences in the way males and females spent their severance pay are that males were more likely to invest the money in a business venture whereas females were more likely to deposit it in their bank account. While the research sample was not a representative random sample, these findings fit within the larger bodies of literature that discuss labor flexibilization, and occupational segregation by gender. Future work that builds on this research might include a similar study of flexible factory workers in Manaus with a statistically
143 significant sample. By collecting data on both flexible and fully contracted factory workers from the s ame firm(s), the researcher could better detail discrepancies in wages and benefits between these two classes of workers. Similar information could be sufficiently disaggregate d by contract type. With such a data set one could also analyze difference between male and female workers in terms of earnings, and distribution among skilled positions. Additionally, one might be able to gain access to such a data set from the Amazonas M etal Workers Union in Manaus.
144 Table 4 1. Respondents' a ge g roups , d isaggregated by g ender and e mployee t ype Age Female Fully contracted Female Flexible Female Distribution Male Fully contracted Male Flexible Male Distribution 19 25 5 2 35% 4 3 36.8% 26 30 5 0 25% 3 0 15.8% 31 40 5 2 35% 3 1 21.1% 41 53 1 0 5% 5 0 26.3% n=39 * Key informants excluded from table structured Interviews
145 Table 4 s elf i dentified c olor/ r ace Color N. Female N. Male Percent Cumulative % Brown 13 17 76.9% 76.9% Black 1 1 5.1% 82.1% White 2 1 7.7% 89.7% Yellow 2 0 5.1% 94.9% Unspecified 2 0 5.1% 100.0% Total 20 19 100.0% N. = 39 * Key informants excluded from analysis Source: structured Interviews
146 Table 4 3. Skilled vs. u nskilled w orkers, d isaggregated by g ender and e mployee t ype Gender Employee Type Unskilled Skilled Total N. % N. % N % Female Fully contracted 9 45% 7 35% 16 80% Flexible 4 20% 0 0% 4 20% Male Fully contracted 9 47% 6 32% 15 79% Flexible 2 11% 2 11% 4 21% Total 24 62% 15 38% 39 100% n= 39 * Key informants excluded from analysis structured Interviews
147 Table 4 4. High s chool g raduates among r espondents, d isaggregated by g ender and e mployee t ype Female Fully contracted Female Flexible Male Fully contracted Male Flexible % High School Graduate* 16 2 15 4 94.9% Non Graduate 0 2 0 0 5.1% Total 16 4 15 4 100.0% n=39 * Includes respondents that have also completed college ** Key informants excluded from analysis structured Interviews
148 Table 4 5. Number of t raining c ertificates, d isaggregated by g ender and e mployee t ype Number of Certificates Fully contracted Flexible Female Male % Female Male % Total 0 2 2 12.9% 0 0 0.0% 13% 1 0 4 12.9% 0 3 37.5% 23% 2 5 2 22.6% 2 1 37.5% 32% 3 or more 2 6 25.8% 2 0 25.0% 32% Total 9 14 74.2% 4 4 100.0% 100% n=31 * Missing responses excluded from analysis Semi structured Interviews
149 Table 4 6. Educational a ttainment t ertiary s chools, d isaggregated by e mployee t ype and g ender Educational Attainment/Progress Fully contracted Flexible N. Female N. Male Incidence N. Female N. Male Incidence No College 8 5 42% 2 2 50% Trade School Finished 2 2 13% 0 1 13% Trade School in Progress 3 1 13% 0 0 0% College Graduate 1 2 10% 0 0 0% College in Progress 2 4 19% 2 1 38% Degree 0 1 3% 0 0 0% n=39 * Key informants excluded from analysis
150 Table 4 7. Living a rrangement, d isaggregated by g ender and e mployee t ype Female Male Living arrangement Fully contracted Flexible % Fully contracted Flexible % Only Adult in Home** 4 0 20.0% 1 0 5.3% Lives with Partner/Spouse 3 3 30.0% 10 2 63.2% Lives with Parents (but not Spouse/Partner) 8 1 45.0% 4 2 31.6% Lives with Roommates 1 0 5.0% 0 0 0.0% Total 16 4 100.0% 15 4 100.0% n=39 * Key informants excluded from table ** may include individuals that have under age children living with them structured Interviews
151 Table 4 8. Number of c hildren among p articipants, d isaggregated by g ender and l iving a rrangement Number of Living Children Only Adult in Home Lives with Spouse/Partner Lives w/ Extended Family (Excluding Spouse/Partner) Lives with Roommate(s) Female Male Female Male Female Male Female Male N % N % N % N % N % N % N % N. % 0 0 0% 1 100% 2 33% 4 33% 5 56% 4 67% 1 100% 0 0% 1 3 75% 0 0% 2 33% 2 17% 3 33% 2 33% 0 0% 0 0% 2 1 25% 0 0% 1 17% 1 8% 1 11% 0 0% 0 0% 0 0% 3 0 0% 0 0% 1 17% 4 33% 0 0% 0 0% 0 0% 0 0% 4 0 0% 0 0% 0 0% 1 8% 0 0% 0 0% 0 0% 0 0% Total N. 4 0 4 8 4 2 0 0 Key informants excluded from analysis structured Interviews
152 Table 4 9. Ages of w orkforce ( e lectronics s ubsector o nly), d isaggregated by g ender M anaus 2010 Age Range Female Male Both N. % N. % Dist.% 14 18 257 2% 202 2% 2% 19 25 3380 29% 3564 31% 30% 26 30 2945 25% 3009 27% 26% 31 40 3313 28% 2896 26% 27% 41 53 1631 14% 1448 13% 13% >53 207 2% 228 2% 2% Total 11733 100% 11347 100% 100% Source: Compiled from Brazilian National Demographic Census (2010)
153 Table 4 10. Composition of the e lectronics m anufacturing w orkforce by m arriage s tatus, d isaggregated by g ender Marital Status Female Male Total Dist. N. % N. % N % Single 8495 75% 8167 72% 16662 73% Married 2894 25% 3178 28% 6072 27% Total 11389 100% 11345 100% 22734 100% Source: Compiled from Brazilian National Demographic Census (2010)
154 Table 4 11. Educational a ttainment t ertiary s chools, d isaggregated by g ender and e mployee t ype Educational Attainment Fully contracted Flexible N. Female N. Male Incidence N. Female N. Male Incidence No College 8 5 42% 2 2 50% Trade Program Incomplete* 3 1 13% 0 0 0% Trade School Graduate 2 2 13% 0 1 13% College Incomplete* 2 4 19% 2 1 38% College Graduate 1 2 10% 0 0 0% Graduate Degree 0 1 3% 0 0 0% Total 16 15 100% 4 4 100% n=39 * "incomplete" indicates that the respondent had complete at least one credit hour of a tertiary program, all though they never graduated. It does not imply that they are currently enrolled. Source: Authors Semi structured Interviews
155 Table 4 12. Educational a ttainment t ertiary s chools, d isaggregated by g ender and e mployee ( s killed w orkers e xcluded from a nalysis) Educational Attainment Fully contracted Flexible N. Female N. Male Incidence N. Female N. Male Incidence No College 6 5 61% 2 2 67% Trade Program Incomplete* 2 1 17% 0 0 0% Trade School Graduate 0 1 6% 0 0 0% College Incomplete* 1 1 11% 2 0 33% College Graduate 0 1 6% 0 0 0% Total 9 9 100% 4 2 100% n= 24** ** Skilled workers excluded from analysis * "incomplete" indicates that the respondent had complete at least one credit hour of a tertiary program, all though they never graduated. It does not imply that they are currently enrolled. structured Interviews
156 Table 4 13. Number of t raining c ertificates, d isaggregated by g ender and e mployee t ype Number of Certificates Fully contracted Flexible Female Male % Female Male % 0 2 2 12.9% 0 0 0.0% 1 0 4 12.9% 0 3 37.5% 2 5 2 22.6% 2 1 37.5% 3 or more 2 6 25.8% 2 0 25.0% Total 9 14 74.2% 4 4 100.0% n=31 * Key excluded informants from analysis structured Interviews
157 Table 4 14. Number of t raining c ertificates ( o nly u nskilled w orkers c onsidered), d isaggregated by g ender and e mployee t ype Number of Certificates Fully contracted Flexible Both Female Male % Female Male % Total Cum % 0 2 2 33.3% 0 0 0.0% 4 17% 1 0 3 25.0% 0 1 16.7% 4 33% 2 2 1 25.0% 2 1 50.0% 6 58% 3 or more 0 2 16.7% 2 0 33.3% 4 75% Total 4 8 100.0% 4 2 100.0% 18 * Key informants excluded from table, missing responses also excluded structured Interviews
158 Table 4 15 . Principal b enefits p articipants r eported r eceiving from t heir e mployers, d isaggregated by e mployee t ype Benefit Fully Contracted Flexible N. Yes Out of N.* % N. Yes Out of N.* % Transportation 31 31 100% 8 8 100% Meals 31 31 100% 8 8 100% Health Plan 31 31 100% 2 8 25% Dental Plan 20 25 80% 2 8 25% Life Insurance 25 25 100% 2 8 25% Discounted Pharmacy 15 21 71% 2 8 25% End of Year Bonus 15 16 94% 1 8 13% * Missing responses excluded from analysis structured Interviews
159 Table 4 16. Mean b ase e arnings, d isaggregated by e mployee t ype ( c urrency BRL$) Employee Type Mean Std. Dev. Fully Contracted $1,259.90 664.09 Flexible $853.71 170.3 n=38* *Informant 31 was excluded from the analysis because she did not work a fully 40 hours ** Key informants excluded from analysis structured Interviews
160 Table 4 17. Mean b ase e arnings of u nskilled w orkers, d isaggregated by e mployee t ype ( c urrency BRL$) Employee Type Mean Std. Dev. Fully Contracted $996.50 709.2 Flexible $957.09 117.4 N=23* *Informant 31 was excluded from the analysis because she did not work a fully 40 hours ** Key informants excluded from analysis structured Interviews
161 Table 4 18. Participants' a ssessment of the Manaus Free Trade Zone Main t hemes e merging in r esponses, d isaggregated by e mployee t ype and g ender Themes Fully contracted Flexible Total N. Female N. Male Incidence N. Female N. Male Incidence % Essential for the population as a source of jobs. 10 7 55% 1 2 38% 51% Companies are wealthy 2 2 13% 1 0 13% 13% The model produces revenue for the government and people 1 1 6% 0 0 0% 5% The model helps preserve the forest 1 0 3% 0 0 0% 3% The model has been a source of considerable migration 1 3 13% 1 0 13% 13% It would be tragic should the model ever be moved from Manaus 1 1 6% 0 1 13% 8% The factories of the district bread injuries and health problems among workers 1 0 3% 1 0 13% 5% Most jobs at the district don't provide security 1 3 13% 2 0 25% 15% Business in the district need to innovate and invest more in technology and products 1 1 6% 0 1 13% 8% There is lack of people qualified to work in the industries at the district 2 0 6% 0 0 0% 5% Jobs are generally of low quality 4 2 19% 2 0 25% 21% Injuries are frequent 2 1 10% 1 0 13% 10% *Key informants excluded; ** multiple responses allowed structured Interviews
162 Table 4 19. Reason for l eaving l ast j ob in the MID, r esponses d isaggregated by g ender and e mployee t ype Condition of Termination of Employment Fully contracted Flexible Total N. Female N. Male Incidence N. Female N. Male Incidence % Quit on Own Account 3 3 46% 1 0 33% 44% Laid off 3 1 31% 0 0 0% 25% Fired 2 0 15% 0 0 0% 13% Unknown 1 0 8% 1 1 67% 19% Total 9 4 100% 2 1 100% 100% n=16* * Only unemployed informants considered structured Interviews
163 Table 4 20. Perceived d ifference in t reatment of f lexible workers c ompared to f ully contracted workers, r esponses d isaggregated by g ender and e mployee t ype ( m ulti answers a llowed) Perceived Difference in Treatment Response Female Male Fully contracted Flexible Fully contracted Flexible Difference in benefits Yes 7 1 3 1 No 0 0 0 0 Unspecified 9 3 12 3 Difference in treatment by supervisors and managers Yes 1 0 3 0 No 8 3 5 2 Unspecified 7 1 7 2 Fully contracted workers discriminate/ belittle them Yes 7 2 4 2 No 2 1 2 0 Unspecified 7 1 9 2 Institution discrimination (i.e. treatment by H.R, access to meal room, uniforms) Yes 3 1 3 1 No 2 1 0 0 Unspecified 11 2 12 3 Blame placed on flexible workers Yes 3 0 1 0 No 1 0 0 0 Unspecified 12 4 14 4 n=39 * Key informants excluded from analysis Source: structured Interviews
164 Table 4 21. Future p lans with r egards to e mployment/ p rofession c oded by e mployee t ype and g ender ( m ulti a nswers a llowed) Future Plans Fully contracted Flexible Dist. Both, % N. Female N. Male Incidence N. Female N. Male Incidence Finnish/school pursue degree 13 9 71.0% 4 3 88% 74% Start own company or pursue self employment 1 2 9.7% 0 1 13% 10% Move out of industrial employment 7 5 38.7% 2 1 38% 38% Work in Industrial District 6 8 45.2% 1 1 25% 41% Move up in company 1 3 12.9% 0 1 13% 13% Maintain current employment and status 0 4 12.9% 0 0 0% 10% Total N =39 * Key informants excluded from analysis structured Interviews
165 Table 4 22. High s chool g raduates among r espondents, d isaggregated by g ender and e mployee t ype Female Male Fully contracted Incidence Fully contracted Incidence High School Graduate* 18 90% 19 100% Non Graduate 2 10% 0 0% Total 20 100% 19 100% n=39 * Includes respondents that have also completed college structured Interviews
166 Table 4 23. Number of t raining c ertificates, d isaggregated by g ender ( b oth s killed and u nskilled w orkers i ncluded) Number of Certificates Female Male Incidence N. Incidence N. 0 15% 2 11% 2 1 0% 0 39% 7 2 54% 7 17% 3 3 or more 31% 4 33% 6 Total 100% 13 100% 18 *Excluded from analysis: missing cases structured Interviews
167 Table 4 24. Are t here m ore m ales or f emales e mployed in y our w ork a rea? Responses d isaggregated by g ender Response Female Male Dist. Both N. Female Incidence N. Male Incidence Dist. More Females 13 65% 5 26% 46% More Males 7 35% 12 63% 49% Equal 0 0% 1 5% 3% No response 0 0% 1 5% 3% Total 20 100% 19 100% 100% * Key informants excluded from analysis structured Interviews
168 Table 4 25. Sex of o verseer, d isaggregated by r espondents' g ender Sex of Overseer N. Female Incidence N. Male Incidence Distribution Women 11 55% 4 21% 38% Man 8 40% 13 68% 54% No response 1 5% 2 11% 8% Total 20 100% 19 100% 100% * Key informants excluded from analysis structured Interviews
169 Table 4 26. Content a nalysis, g endered t hemes m ost a ddressed in i nterview, d isaggregated by i ndividual r espondents ( m ulti a nswers a llowed ) 1 Gender Related Theme/Comment N. Female N. Male % of Total Participants Who Mentioned Theme Women have nimble fingers, greater manual dexterity 4 2 13% Women concentrated in manufacturing 4 5 20% Females are caddy or jealous 1 1 4% Men concentrated among machine operators 5 2 16% Men are Physically Stronger than Women 3 4 16% Men are brutish (clumsy) 2 1 7% Increase in female participation in certain areas 3 0 7% Increase in Male Participation in Certain Areas 1 0 2% Women concentrated in clerical positions 1 6 16% Women in Leadership 5 0 11% Men in leadership 2 2 9% Total 31 23 * Key informants included in analysis 1. In some cases these themes appeared in response to specific questions, where as in other cases they emerged organically throughout the course of the interview. In some cases I used the "tell me why you think that is" probing technique, depending on the nature of the discussion. struc tured Interviews
170 Table 4 27. Future p lans with r egards to e mployment/ p rofession c oded by g ender and e mployee c ategory ( m ulti a nswers a llowed) Future Plans Female Male Fully contracted Flexible % Fully contracted Flexible % Finnish/school pursue degree 13 4 85.0% 9 3 63.2% Start own company or pursue self employment 1 0 5.0% 2 1 15.8% Move out of industrial employment 7 2 45.0% 5 1 31.6% Work in Industrial District 6 1 35.0% 8 1 47.4% Move up in company 1 0 5.0% 3 1 21.1% Maintain current employment and status 0 0 0.0% 4 0 21.1% Total 28 7 100.0% 31 7 100.0% n = 39 * Key informants excluded from analysis structured Interviews
171 Table 4 28. Marital s tatus of r espondents, d isaggregated by g ender Marital Status N. Female Incidence N. Male Incidence Single 14 70% 12 63% Married 4 20% 7 37% Divorced/Separated 2 10% 0 0% Total 20 100% 19 100% * Key informants excluded from analysis structured Interviews
172 Table 4 29. Living a rrangement, d isaggregated by m arital s tatus and g ender Living Situation Female Male Both Single % Married % Divorced % Single % Married % Divorced % Total On ly Adult in Home** 3 21% 0 0% 1 50% 1 8% 0 0% 0 0% 12.8% Lives with Partner/Spouse 2 14% 4 100% 0 0% 5 42% 7 100% 0 0% 46.2% Lives with Parents (but not Spouse/Partner) 8 57% 0 0% 1 50% 6 50% 0 0% 0 0% 38.5% Lives with Roommates 1 7% 0 0% 0 0% 0 0% 0 0% 0 0% 2.6% Total 14 70% 4 20% 2 10% 12 63% 7 37% 0 0% 100.0% n=39 * Key informants excluded from table ** may include individuals that have under age children living with them Compiled from research sample
173 Table 4 30. Who is p rimarily r esponsible for p aying the g iven e xpense? D isaggregated by g ender and l iving a rrangements Expense Person Primarily Responsible for Paying Expense 1 Living Arrangement Only Adult in Home ** Spouse/Partner 2 Extended Family, Not Spouse/Partner Roommates Female Male Female Male Female Male Female N. % N. % N. % N. % N. % N. % N. % Rent/ Mortgage Other Person 1 25% 1 100% 0 0% 1 8% 1 11% 0 0% 0 0% Divided Equally 0 0% 0 0% 3 50% 0 0% 0 0% 0 0% 1 100% Self 3 7 5% 0 0% 0 0% 4 33% 2 22% 0 0% 0 0% N/A 0 0% 0 0% 3 50% 7 58% 6 67% 6 100% 0 0% Total* 2 0% 5% 30% 63% 45% 32% 5% Groceries Other Person 0 0% 0 0% 1 17% 2 17% 3 50% 1 20% 0 0% Divided Equally 0 0% 1 100% 5 83% 3 25% 1 17% 2 40% 1 100% Self 4 100% 0 0% 0 0% 6 50% 5 83% 1 20% 0 0% N/A 0 0% 0 0% 0 0% 1 8% 0 0% 1 20% 0 0% Total* 20% 6% 30% 67% 45% 5% 5 % *M issing cases excluded from the analysis 1. Refers to the person who uses their own earnings to pay for the expense 2. Includes individuals that live with their partner and extended family 3. includes tuition expenses for children, and or self ** may include individuals that have under age children living with them structured Interviews
174 Table 4 30 . Expense Person Primarily Responsible for Paying Expense 1 Living Arrangement Only Adult in Home ** Spouse/Partner 2 Extended Family, Not Spouse/Partner Roommates Female Male Female Male Female Male Female N. % N. % N. % N. % N. % N. % N. % Clothing, Toys Accessories for Children in the Household Other Person 0 0% 0 0% 1 17% 2 33% 1 11% 0 0% 0 0% Divided Equally 0 0% 0 0% 1 17% 0 0% 0 0% 0 0% 0 0% Self 2 50% 0 0% 3 50% 6 50% 3 33% 2 33% 0 0% N/A 2 50% 1 100% 1 17% 4 67% 5 56% 4 67% 1 100% Total* 20% 5% 30% 63% 40% 32% 5% *Missing cases excluded from the analysis 1. Refers to the person who uses their own earnings to pay for the expense 2. Includes individuals that live with their partner and extended family 3. includes tuition expenses for children, and or self ** may include individuals that have under age children living with them structured Interviews
175 Table 4 31. Number of c hildren among p articipants, d isaggregated by g ender and l iving a rrangement Number of Living Children Only Adult in Home Lives with Spouse/Partner Lives w/ Extended Family (Excluding Spouse/Partner) Lives with Roommate(s) Female Male Female Male Female Male Female Male N % N % N % N % N % N % N % N % 0 0 0% 1 100% 2 33% 4 33% 5 56% 4 67% 1 100% 0 0% 1 3 75% 0 0% 2 33% 2 17% 3 33% 2 33% 0 0% 0 0% 2 1 2 5% 0 0% 1 17% 1 8% 1 11% 0 0% 0 0% 0 0% 3 0 0 % 0 0% 1 17% 4 33% 0 0% 0 0% 0 0% 0 0% 4 0 0% 0 0% 0 0% 1 8% 0 0% 0 0% 0 0% 0 0% Total N. 4 0 4 8 4 2 0 0 Key informants excluded from analysis structured Interviews
176 Table 4 32. Participants with c hildren l iving in the h ome, d isaggregated by g ender and l iving a rrangement Number of Dependent Minors in Household Only adult in home Spouse/Partner Extended Family/ No Spouse Female Male Female Male Female Male N. % N. % N. % N. % N. % N. % 1 2 67% 0 0% 2 50% 3 75% 3 100% 1 100% 2 1 33% 0 0% 0 0% 1 25% 0 0% 0 0% 3 0 0% 0 0% 2 50% 0 0% 0 0% 0 0% Total 3 0 4 4 3 1 *Key informants excluded from analysis structured Interviews
177 Table 4 33. After y ou h ave p aid all h ousehold b ills and e xpenses , * h ow m uch of y our e arnings are l eft for y our p ersonal u se?* Gender % Mean N. Std. Dev. Female 14% 9 0.115 Male 23% 5 0.156 Total 16% 14 0.133 * Includes expenses associate with children ** Outlying case excluded *** Informant 31 was excluded from the analysis **** Key informants excluded from analysis structured Interviews
178 Table 4 34. Average g ross m onthly e arnings w orkers with c hildren in the h ome, d isaggregated by g ender ( c urrency BRL$) Gender Av. Earnings As % Av. Male Earnings N. Std. Dev. Female 1297.11 96% 9 679.0 Male 1356.60 100% 5 490.4 Total 1318.35 97% 14 598.8 * Informant 31 was excluded from the analysis structured Interviews
179 Table 4 35. Average g ross m onthly e arnings u nskilled w orkers with c hildren in the h ome, d isaggregated by g ender ( c urrency BRL$) Gender Av. Earnings As % Av. Male Earnings N. Std. Dev. Female $979.66 86% 9 265.9 Male $1140.25 100% 5 92.7 Total $1043.90 92% 14 221.4 * Informant 31 was excluded from the analysis Semi structured Interviews
180 Table 4 36. How d id y ou u se y our s everance p ay? Disaggregated by g ender and e mployee t ype ( o nly o ne a nswer c onsidered) How Severance Pay Was Spent Female Male Fully contracted Flexible % Fully contracted Flexible % Savings 2 0 15% 0 0 0% Paid off debts 4 1 38% 3 0 38% Housing 1 1 0 8% 1 0 13% Education Expenses 2 1 0 8% 1 0 13% Spent on Children 2 0 15% 0 0 0% Travel 0 2 15% 1 0 13% Business investment 3 0 0 0% 2 0 25% Total 10 3 100% 8 0 100% 1. Building or remodeling a home, advancing construction 3. Went towards purchasing merchandise, facilities, or savings for a business venture. * Key informants excluded from analysis Source: structured Interviews
181 CHAPTER 5 FINAL CONSIDERATIONS After nearly half a century, the Manaus Free Trade Zone continues with an evident vigor, as demonstrated by the unprecedented revenues and production levels recorded by Suframa over the last four years. While individual businesses in its industrial district have come and gone with changes in technology, markets, and macro economic policy, the model itself, has survived a number of crises. 1 Although foreign investment continues to pla project, as many of the largest and most successful firms in its industrial district are foreign companies, the model is largely inward looking, and reminiscent of the ISI paradigm. Even though most re venues generated by businesses in the Manaus Industrial District come from the domestic consumer market, history has shown that the model remains indirectly vulnerable to fluctuations in international markets. Considering omy remains heavily dependent on the export of primary materials to foreign markets, the decline in international demand for Brazilian commodities since 2012 may not brood well for the Manaus Industrial District in the near future. The success of business es in the Manaus Industrial District in recent years reflects the robustness of the Brazilian consumer market in general, characterized by economic and political factors. Brazil ian economists have written extensively on these factors, including the growth of credit services, and income redistribution programs 1 . global financial crisis (2008) n egatively impacted industries in Manaus. In 2009 revenues generated by businesses in the Manaus Industrial District declined by 13% from the previous year, with the workforce likewise declining by 13%.
182 which grant the lower and working classes greater purchasing power (Della Colleta, 2013; Fe ltrim, 2009) . Although this trend in consumer spending habits has proven profitable for businesses in the Manaus Industrial District, over the last decade, the growth in demand for goods produced in Manus is unlikely to continue at present levels (Cigana, 2014) . The relative success of businesses in the Manaus Free Trade Zone over the last decade is more indicative of national economic conditions, than increased innovation on the part of producers and researchers. Both critics and proponents of the model suggest t he need for companies in the Manaus Industrial District to invest in new technology and research in order to create more competitive, value added products. They argue that product innovation would increase long term viability of the model reducing dependen cy on tax incentives. They suggest that the fiscal haven created by the Manaus Free Trade Zone would provide an ideal environment for research and the development of cutting s natural resources (Bonfim, 2006) . In their view such a direction would convert Manus into an epicenter of sustainable development in the nation (Pereira 2006). ederal tax code, and opposition from industrialists in other parts of Brazil that promote greater liberalization of the national economy, the Manaus Free Trade Zone is unlikely to disappear anytime soon. Politicians and policy makers are conscious that any attempt to reduce the scope of the project would have severe consequences for the economy and population of Manaus. As much as one percent of the national population resides in Manaus, and the welfare of this important voting block is of the utmost concer n to
183 politicians at the state and federal levels. Additionally, environmental concerns at the national and international level feature prominently in public discourse about the model, since it provides a strong economic alternative to the exploitation and destruction of the Amazon rainforest by local populations. The vocal majority in favor of the preservation Manaus provides a fascinating context for the study of development and related themes, such as restructuring, state led industrialization, and gender, because of its free trade zone. Motivated by the current understanding of development in the literature as including the growth of opportunity and greater equality for historic ally marginalized groups, I have sought to contribute to the understanding of how the gendered composition of the workforce in the Manaus Industrial District has changed in light of macroeconomic trends. As in other manufacturing economies throughout the w orld, women have historically made up an important part of the workforce in Manaus. Through the quantitative analysis of census data I was able to show the changes in the 201 0. There was a pronounced decline in the share of jobs occupied by women after 1991, which roughly corresponds to the opening of the Brazilian economy and the onset of restructuring. During this time, many businesses in the industrial district experienced a crisis of sorts, as demand for their products plummeted. Most businesses reacted by reducing their total number of employees and many were forced to shut down operations. The marked decline in the share of women employed in the Manaus Industrial District between 1991 and 2000 is a robust example of defeminization of the workforce, not unlike that observed in other contexts (Pearson 1990).
184 Although the overall share of females employed in manufacturing declined with the opening of the Brazilian economy, t he incidence of women in skilled positions suffered much less than those in unskilled labor intensive jobs, and even increased in some subsectors such as electronics manufacturing. Although the growth of female participation in such positions initially see ms to contradict the literature on gender and restructuring, which suggest that skilled positions tend to favor male over female workers, this trend in Manaus is understandable when we consider the respective demographic characteristics of the male and fem ale population in Manaus and Brazil in general. As female educational attainment rates continue to outpace that of males alongside increases in the overall economic participation rate of women, it is reasonable to think that this trend towards more women w orkers in skilled positions in the Manaus Industrial District will persist in coming years. This finding regarding the growth of the share of women in skilled positions in the electronics subsector in 2000 and 2010, was further substantiated by qualitative interviews with individuals who work in the Manaus industrial District, who expressly indicated they had seen the growth of women in such positions in recent years. quality of l ife, inequality in the labor market between men in women should be of the utmost concern to individuals and institutions concerned with gender equality. Much of the observable inequality between male and female workers in terms of wages is related to occup ational segregation, although the literature also suggests a number of other factors at play, including a so an unofficial barrier impeding and different decision making tendencies between
185 men and w omen likely rooted in socialization (Croson, 2009; Davidson, 1992) . Researchers in the United States have provided evidence that women even at the highest echelons of management often accept pay inferior to men. Althou gh I am unaware of any studies examining this question in the context of the Manaus Free Trade Zone, it seems likely that such disparities also exist between male and female managers in the Manaus Industrial District. In my analysis of census data a found that the wage gap between males and females in the Manaus Industrial District has followed trends similar to those recorded elsewhere. While disparity in wages was particularly pronounced in 1970, by 2000 women workers earned on average 75% of what males earned, a significant improvement over previous decades (45% in 1970). While these findings suggest a positive move in the direction of greater gender equality, they indicate that there is yet much progress to make. The most conspicuous element of restru cturing in the Manaus Industrial District was manifest in the pronounced reduction of the workforce following the opening of the Brazilian economy. For nearly a decade thereafter, the workforce remained significantly te significant increases in production and revenue. Slower job growth in the district alongside increased productivity and investment, suggests the substitution of automated and capital intensive processes for manual labor, which according to the literatur e is symptomatic of restructuring. 2 A lso symptomatic of restructuring is the increased use of flexible forms of labor. 2 . Suframa provides investment figures for the Manaus Industrial District between 2000 and 2013, indicating the annual aggregate amount in US dollars invested by businesses in their own operations. According to these figures between 2000 and 2008 investment increased by an average of 23% per year, whereas ag gregate revenues increased by an average of 19% per year. In comparison the workforce grew by an average of 10% per year (Suframa, 2010: 25, 71, 89).
186 According to Suframa , flexible workers as a percentage of the total workforce in the Manaus Industrial District have fluctuated between 6% and 12% annually, with increased use during seasons of peak production. Regretfully , few studies address this class of workers which are an integral part of the labor order i n the Manaus Free Trade Zone. This research contribute s to the literature on labor flexibility in the Manaus Free Trade Zone by sheddin g light on the characteri stics and perceptions of this important group of workers . While I expected to find notable differences between flexible and fully contracted worker s in terms of their demog raphic a nd professional characteristics, the differences were less pronounced than anticipated. I found negligible differences between flexible and fully contracted worker s in terms of their educational attainment and professional qualifications. Although both flexible and fully contracted worker s readily acknowledged the latter are no more often occupied by finding is based on the comparison of research participants from different contract types, as well as the insights of the key informants that recruit and oversee industrial labor in the district. The main differences observed between flexible and fully contracted worker s w ere in terms of the benefits they receive, with the latter generally receiving superior benefits and higher wages. These findings are based on a comparison of their income data as well as the experiences shared by those interviewed. When we consider the p ronounced difference in treatment flexible and fully contracted worker s receive from their employers, despite minimal differences between both classes of workers in terms of their qualifications, such a contrast borders on
187 injustice. It suggests that unsk illed workers remain largely disempowered over their employment prospects in manufacturing, since their professional qualifications and experience weigh little in determining their contract type. Fortunately, however, there is evidence that the use of flex ible workers in the Manaus Industrial District is in decline. According to key informants in the research study, the pronounced inequality between workers based on their contract type is being ameliorated through legislation and union demands, among the m ost notable of which is the requirement that factories provide health plans for their workers regardless of contract type. The Manaus Free Trade Zone is likely to remain of interest to students and scholars of development, as long as it exists. With the s uccessful approval of constitutional amendment 506/10, the fiscal incentives enjoyed by business in the Manaus Industrial District are now schedule to last until 2073 (Nascimento, 2014) . How the workforce at the Manaus Indus trial District evolves in coming decades, particularly with advances in technology and future adjustments in the federal tax code, will occupy generations of researchers to come. In assessing whether or not the Manaus Free Trade Zone has succeeded in fulfi lling the purpose for which it was created, one must carefully consider the very definition of development. While the project has brought unprecedented growth to the population and economy of Manaus, and contributed to a number of public infrastructure pro jects through tax revenues, it has arguably been less successful in increasing the standard of living and quality of life of countless individuals in the lower tiers of factory employment. Businesses in the Manaus Industrial District are essentially no dif ferent from corporations elsewhere, who seek first and foremost profits in the
188 interest of their shareholders. Such a paradigm inevitably leads to the exploitation of the local workforce, as manifest particularly in the use of flexible workers , stress rela ted injuries, and high turnover rates.
189 LIST OF REFERENCES Alice, A. (2013, September 6). ParticipaÃ§Ã£o da indÃºstria no PIB pode cair para 9,3% em 2029, aponta estudo da Fiesp. SÃ£o Paulo, SP. Alper, A. (2012, Nov 13). Brasil e Argentina tÃªm maior carga tributÃ¡ria da AmÃ©rica Latina, diz OCDE. EstadÃ£o.com.br . Amanatullah, E. a. (2010). Negotiating gender roles: gender differences in assertive negotiating are mediated by women's fear of backlash and attenuated when negotiating on behalf of others. Journal of personality and social psychology, 98 (2), 256 267. Andrade, A. J. (2012, September 26). Costs of Doing Business in Brazil. Gainesville, Fl, USA: Center for Latin American Studies and Warrington School of Business. Atkinson, R. a. (2001, Summer). Social Research Update (33). Balestro, M. V. (2011, May). Seguro desemprego no Brasil: a possibilidade de combinar proteÃ§Ã£o social e melhor funcionamento do mercado de trabalho. Sociedade e Estado , 185 208. Barbosa, I. (2007). ChÃ£o de FÃ¡brica: Ser Mulher OperÃ¡ria no PÃ³lo Industrial de Manaus. Manaus, AM, Brazil: Editora Valer. BÃ¡rcena, A. (2010). What kind of State? What kind of equality? BrasÃlia: United Nations. Basit, T. N. (2003, Summer). Manual or Electronic? The role of coding in qualitative data analysis. Educational Researc, 45 (2), 143 154. BenerÃa, L. (2003). Gender, Development, and Globalization: Economics as if all People Mattered. New York: Routledge. B enerÃa, L. a. (1987). The Cross Roads of Class and Gender: Industrial Homework, Subcontracting and Household Dynamics in Mexico City. Chicago: University of Chicago Press. Bernard, H. R. (2000). Social Research Methods: Qualitative and Quantitative Approac hes. Thousand Oaks, CA, USA: Sage Publications Inc. Bertelli, S. M. (2014, January 17). TerceirizaÃ§Ã£o: forma de gestÃ£o de trabalho ou meio de acomodaÃ§Ã£o da produÃ§Ã£o capitalista. Retrieved from AssociaÃ§Ã£o Nacional dos Magistrados da JustiÃ§a do Trabalho: htt p://www.anamatra.org.br/index.php/artigos/terceirizacao forma de gestao de trabalho ou meio de acomodacao da producao capitalista BÃ©rtola, L. B. (2012). The Economic Development of Latin America since independence. Oxford, UK: Oxford University Press.
190 Bizz otto, M. (2014, 24 2014). Dilma 'estranha' aÃ§Ã£o da UE na OMC contra a Zona Franca de Manaus. BBC/Brasil . London, UK: BBC. Bonfim, R. a. (2006). Zona Franca de Manaus: condicionantes do futuro. Manaus: Editora Valera. Borsato, F. P. (2005). A Seguridade Soc ial no Brasil: Um desafio para o nÃºcleo de seguridade social e polÃtica social de saÃºde da Unioeste. 2Â° SeminÃ¡rio Nacional Estado e PolÃticas Sociais no Brasil. Cascavel: Universidade Estadual do Oeste de ParanÃ¡. Boserup, E. (1986). Woman's role in economi c development. Brookfield: Gower. Botelho, A. J. (2006). Redesenhando o Projeto Zona Franca de Manaus. Manaus, AM, Brazil: Editora Valer. Bueno, E. (2005). Brasil: uma HistÃ³ria a IncrÃvel Saga de um PaÃs. SÃ£o Paulo: Editora Ãtica. Bulmer Thomas, V. (2003). The Economic History of Latin America Since Independence (2nd ed.). New York, NY: Cambridge University Press. Campos, A. G. (2011). Direito ao Trabalho: ConsideraÃ§Ãµes Gerais e Preliminares. BrasÃlia: IPEAD. Caraway, T. L. (2007). Assembling Women. New Yor k: Cornell University. Cardoso, F. H. (1977). AmazÃ´nia: expansÃ£o do capitalismo. SÃ£o Paulo: Editora Brasiliense. Carvalho, G. M. (2011). Sobrecarga de trabalho e Adoecimento no Polo Indusrial de Manaus. Psicologia em Revista., 17 (3), 465 482. Chant, S. &. (2012). Fixing women or fixing the world? 'Smart economics', efficiency approaches, and gender equality in development. Gender & Development , 517 529. Chant, S. (2003). Gender and Employment. In S. Chant, Gender and Latin America (pp. 194 227). London: Lat in American Bureau. Cigana, C. (2014, May 30). AvanÃ§o de apenas 0,2%. ZH Economia . Retrieved May 30, 2014, from http://zh.clicrbs.com.br/rs/noticias/economia/noticia/2014/05/consumo e investimento desaceleram e economia perde folego 4514731.html Cleary, D. (2004). Lost Altogether to the Civilised World: race and the Cabanagem in Northern Brazil, 1750 to 1850. Comparative Studies in Society and Histority, 40 (1), 109 135.
191 Costa, M. d. (2005). O Sistema de RelaÃ§Ãµes de Trabalho no Brasil: alguns traÃ§os histÃ³ricos e sua precarizaÃ§Ã£o atual. RBCS, 20 (59), 111 170. Croson, R. &. (2009). Gender Differences in Preferences. Journal of Economic Literature , 448 474. Cruz, L. (2011, Ju ly 15). Zona Franca de Manaus bate recorde de faturamento nos primeiros cinco meses do ano. Dalal, K. (2011, January). Does economic empowerment protect women from intimate partner violence? Journal of Injury and Violence , 35 44. D'Almeida, C. G. (1982). Z ona Franca: De Onde, Para Onde. Manaus: Imprensa Oficial do Estado do Amazonas. Davidson, M. (1992). Shattering the glass ceiling: The woman manager. London : Paul Chapman. de Souza Gomes, M. M. (2011). O Mundo do Trabalho no Cais do Porto da Manaus Moderna: O Carregador de bagagens e o trabalho precÃ¡rio. II ConferÃªncia do Desenvolvimento , 1 21. IPEA. Della Colleta, R. (2013, 2013 Oct.). Ipea: Bolsa FamÃlia gera o maior efeit o sobre consumo. EstadÃ£o: Economia e NegÃ³cios , p. 15. Retrieved April 4, 2014, from http://economia.estadao.com.br/noticias/economia geral,ipea bolsa familia gera o maior efeito sobre consumo,167482,0.htm DesprÃ©s, L. A. (1991). Manaus: Social Life and Work in Brazil's Free Trade Zone. Albany, NY, USA: State University of New York Press. Diamond, L. (1994). Introduction: Political Culture and Democracy. In L. Diamond, Political Culture and Democracy in Developing Countries (pp. 1 27). Boulder & London: Lynne Rienner Publishers. do EspÃrito Santo, S. (2013, September 25). No Amazonas, Sindicato dos MetalÃºrgicos deflagra campanha pelo direito Ã creche . (CUT, Producer) Retrieved March 23, 2014, from Central Ãšnica do Trabalhadores: http://www.cut.org.br/destaque central/53310/no amazonas sindicato dos metalurgicos deflagra campanha pelo direito a creche en Druck, G. (2011). Trabalho, PrecarizaÃ§Ã£o e ResistÃªncias: novos e velhos desafios? Caderno CRH, 24 (special 01), 35 57. nd Economic Developmen. Working Paper 17702 . Retrieved from http://www.nber.org/papers/w17702
192 Feltrim, L. E. (2009). Projeto InclusÃ£o Financeira: Perspectivas e Desafios para InclusÃ£o Financeira no Brasil: visÃ£o de diferentes atores. BrasÃlia: Banco Centra l do Brasil. Retrieved from Banco Nacional de Desenvolvimento SustentÃ¡vel. Ferreira, M. A. (2003). O desenvolvimento do Capitalismo em Manaus. Manaus, AM, Brazil: Editora da Universidade Federal do Amazonas. FIESP. (2012). Capital Humano Departamento de AÃ§ Ã£o Regional: Estado de SÃ£o Paulo MÃ©dio Salarial por Setor. Retrieved from FIESP: http://apps.fiesp.com.br/regional/dadossocioeconomicos/informacoessetor.aspx? t=3 Filho, G. (2006). CooperaÃ§Ã£o em Clusters: Perspectivas para o PÃ³lo Industrial de Manaus. BrasÃ lia: Paralelo 15. Fonseca, I. (2013, August 22). Polo Industrial de Manaus e as condiÃ§Ãµes de trabalho. "SituaÃ§Ã£o aterradora": Entrevista especial com Ilan Fonseca . SÃ£o Leopoldo, RS, Brazil. Freitas, C. E. (1999). A Pesca e as PopulaÃ§Ãµes Riberinhas da Amaz Ã´nia Central. Journal de Ecologia , 32 36. Gaspar, L. (2003, July 24). O Nordeste do Brasil . Retrieved from FundaÃ§Ã£o Joaquim Nabuco: http://basilio.fundaj.gov.br/pesquisaescolar./index.php?option=com_content&vie w=article&id=197&Itemid=193 GoldÃn, L. R. (201 1). Labor Turnover Among Maquiladora Workers of Highland Guatemala: Resistance adn Semiproletarianization in Global Capitalism. Latin American Research Review, 46 (3), 133 156. GonÃ§alves, O. O. (2012). Os incentivos TributÃ¡rios na Zona Franca de Manaus e o DesequilÃbrio Concorrencial no Setor de Refrigerantes. Economic Analysis of Law Review, 3 (1), 71 94. Hein, C. (1986). The feminisation of Industrial Employment of Mauritius: A Case of Sex Segregation. In R. a. Hein (Ed.), Sex Inequalities in Urban EMployme nt in hte Third World (pp. 277 312). New York, NY, USA: St Martin's Press. Hemming, J. (2008). Tree of Rivers. London: Thames and Hudson. Humphrey, J. (1987). Gender and Work in the Third World. New York, NY: Travistock Publications. IBGE. (2014). Instituto Brasileiro de Geografia e EstatÃsticas . Retrieved from ibge.gov.br: http://www.ibge.gov.br/home/estatistica/indicadores/trabalhoerendimento/pme/p memet2.shtm
193 Jannuzzi, P. d. (2003). Os Quesitos de MÃ£o de Obra nos Censos DemogrÃ¡ficos Brasileiros de 1960 a 2000. Encontro transdisciplinar espaÃ§o e populaÃ§Ã£o. Kabeer, N. (1999). Resources, Agency, Achievement: Reflections on the Measurement of Women's Empowerment. Development and Change , 435 464. Kingstone, P. (2011). The Political Economy of Latin Amer ica. New York, NY: Routledge. LDCS, S. O. (2014). UN sustainable development indicators. Retrieved June 1, 2014, from www.un.org: http://www.un.org/esa/sustdev/natlinfo/indicators/methodology_sheets/global_ec on_partnership/imports_from_devcountries.pdf Lei te, M. d. (1997). O Trabalho em Movimento. SÃ£o Paulo, Brazil: Papirus Editora. Lemos, J. d. (2007). Mapa de ExclusÃ£o Social no Brasil: Radiografia de um PaÃs Assimetricamente Pobre. Fortaleza: Banco do Nordeste. Lima, D. a. (2012). Mapa do Desemprego no Br asil. SÃ£o Paulo: AgÃªncia do Globo. Retrieved February 15, 2014, from http://oglobo.globo.com/termos de uso/ Lima, V. S. (2009, July). Produzidas no Polo Industrial de Manaus: ExperiÃªncia de Mulheres trabalhadoras. Retrieved March 23, 2014, from Biblioteca Digital de TÃ©ses e DissertaÃ§Ãµes: http://bdtd.bce.unb.br/tedesimplificado/tde_busca/arquivo.php?codArquivo=5793 Maciel, M. S. (2009). DependÃªncia de TrajetÃ³ria nos incentivos fiscais: fragmentaÃ§Ã£o do empresariado na reforma tributÃ¡ria. Retrieved March 10, 2013, from CÃ¢mara dos Deputados: Biblioteca Digital CÃ¢mara: Mahar, D. J. (1976). Fiscal Incentives and Economic Development of Western AmazÃ´nia. Brazilian Economic Studies , 147 174. Mar iotti, H. (2014, March). O complexo de inferioridade do brasileiro. Revista BSP . Martins, R. (2014, Jan 17). FGTS: tire suas dÃºvidas sobre o Fundo de Garantia de Tempo de ServiÃ§Ã³. Empresa Brasileira de ComunicaÃ§Ã£o . Moghadam, V. M. (1999). Gender and the Gl obal Economy. In J. T. N Roberts (Ed.), The Globalization and Development Reader (pp. 135 148). Malden, MA: Blackwell Publishing. Moraes, R. D. (2008). Prazer Sofrimento e SaÃºde no Trabalho com AutomaÃ§Ã£o: Estudo com operadores em empresas japonesas no PÃ³lo Industrial de Manaus. BelÃ©m, ParÃ¡, Brasil.
194 Morandi, L. (2003). Estoque de Capital Fixo no Brasil, 1950 2002. Rio de Janeiro: IPEA. Morgan, D. L. (2008). Snowball Sampling. In L. M. Given, The Sage Encycolopedia of Qualitative Research Methods (pp. 816 817 ). Thousand Oaks, CA, USA: Sage Publications Inc. MS/SAS/DAB/CGAN. (2014). RelatÃ³rio Geral de sÃ©rie histÃ³rica do percentual de acompanhamento das famÃlias em formato CSV. MinistÃ©rio da SaÃºde. Retrieved June 2014, 28, from http://bolsafamilia.datasus.gov.br /w3c/bfa_relconsol.asp Nascimento, L. (2013, June 27). ManifestaÃ§Ã£o contra corrupÃ§Ã£o e pela reforma polÃtica Ã© dispersada pela PolÃcia Militar do DF. Empresa Brasil de ComunicaÃ§Ã£o . Nascimento, L. (2014, March 19). PEC da Zona Franca de Manaus Ã© aprovada em 1Âº turno na CÃ¢mara. Empresa Brasileira de ComunicaÃ§Ã£o . Nunes de Miranda. (2013). ZONA FRANCA DE MANAUS : Desafios e Vulnerabilidades. NÃºcleo de Estudos e Pesquisas do Senado. Senado Federal. Oliveira, A. A. (2007). O perfil do engenheiro da indÃºstica elet rÃ´nica da zona Franca de Manaus; Caracteristicas e transformaÃ§Ãµes. Revista de Biologia e CiÃªncias da Terra, 7 , 183 193. Passarinho, N. (2012, April 10). CCJ aprova prorrogaÃ§Ã£o de incentivos para Zona Franca de Manaus. G1 PolÃtica . Pearson, R. (1986). Latin American Women adn the New INternational Division of Labour: A Reassessment. Bulletin of Latin American Research, 5 (2), 67 79. Pearson, R. (1990). Male Bias and Women's Work in Mexico's Border Industries. Male Bias in the Developm ent Process , 133 163. Pereira, D. (2005). AmazÃ´nia InsustentÃ¡vel. Manaus: Editora Valer. Perkin, H. (1996). The Third Revolution. Professional Elites in the Modern World. London: Routledge. Pochmann, M. (2009). What Brazil Learned from Labour Flexibilizati on in the 1990s. International Labour Review, 148 (3), 269 282. Ramos, A. L. (1997). AcumulaÃ§Ã£o FlexÃvel e Direito do Trabalho. Revista de CiÃªncias Humanas, 15 (22), pp. 76 89. Retrieved from Ãrea de Ciencias, Sociedade e Cultura. Ravena, N. (2005). O abaste cimento no SÃ©culo XVIII no GrÃ£o ParÃ¡. Novos Cadernos NAEA, 8 (2), 125 149.
195 Ribas, R. a. (2011). Is the effect of conditional transfers on labor supply negligible everywhere? University of Illinois. Roberts, J. T. (2003). Trouble in Paradise. New York, NY: R outledge. Roett, R. (2010). The New Brazil, Harrisonburg. Harrisonburg: R.R. Donnelley. Sader, E. (2010, Feb. 11). Brasil, de GetÃºlio a Lula. Brasil Passado e Futuro , 11 29. Sant'Anna, L. (2013, March 31). Desafios da IndÃºstria: Impostos e margens explicam preÃ§os altos. O Estado de SÃ£o Paulo, 31 . Santo, S. d. (2013, September 25). No Amazonas, Sindicato dos MetalÃºrgicos deflagra campanha pelo direito Ã creche . (CNM/CUT, Producer) Retrieved March 23, 2014, from Central Ãšnica dos Trabalhadores: http://www.cut .org.br/destaque central/53310/no amazonas sindicato dos metalurgicos deflagra campanha pelo direito a creche en Scherer, E. F. (2004). Desemprego, trabalho precÃ¡rio e des cidadanizaÃ§Ã£o na Zona Franca de Manaus. Somanlu, 4 (1), pp. 125 145. Schmink, M. &. ( 1992). Contested Frontiers in AmazÃ´nia. New York: Columbia University Press. Schneider, C. A. (2005). Planejamento BÃ¡sico de um Arranjo Produtivo Local voltado para IndÃºsria EletroeletrÃ´nica e Desenvolvimento da IndÃºstria MicroeletrÃ´nica. FundaÃ§Ã£o Certi. FlorianapÃ³lis: Governo do Estado de Santa Catarina. Secretaria de PolÃticas para as Mulheres. (2013, September 20). 20/09 Bolsa famÃlia completa 10 anos com 93% de mulheres titulares . Retrieved from Secretaria de PolÃticas para as Mulher es: http://spm.gov.br/noticias/ultimas_noticias/2013/09/20 09 bolsa familia completa 10 anos com 93 de mulheres titulares SerÃ¡fico, J. &. (2005). A Zona Franca de Manaus e o capitalismo no Brasil. Estudos AvanÃ§ados, 19 (54), 99 113. SerÃ¡fico, M. (2011). Glo balizaÃ§Ã£o e empresariado : estudo sobre a Zona Franca de Manaus. SÃ£o Paulo: Annablume. Sindicato dos MetalÃºrgicos de SÃ£o Paulo e Mogi das Cruzes. (2009, Nov. 13). HistÃ³ria Uma tradiÃ§Ã£o de lutas e conquistas . Retrieved Feb. 28, 2014, from Sindicato dos Me talÃºrgicos de SÃ£o Paulo e Mogi das Cruzes: http://www.metalurgicos.org.br/historia.asp Smith, P. H. (2009). Democracy in Latin America. New York: Oxford University Press.
196 Social, S. d. (2001). Comportamento Social e Trabalhista RelatÃ³rio Geral de ObservaÃ§Ã£ o: Moto Honda da AmazÃ´nia. ObservatÃ³rio Social. FlorianÃ³polis: Sede do ObservatÃ³rio Social. Souza, A. M. (2010). A IntegraÃ§Ã£o da AmazÃ´nia por meio de Incentivos Fiscais Administrados pela Suframa: Reflexos para o desenvolvimento nacional. Manaus: MinistÃ©rio do Desenvolvimento, IndÃºstria e ComÃ©rcio. Retrieved November 13, 2014, from http://www.sae.gov.br/seminarioamazonia/wp content/uploads/2010/08/Artigo SUFRAMA.pdf Souza, W. J. (1998). Fordism and its multiple sequels: the re organization of work in Br itain, France, Germany and Japan . Retrieved 03 19, 2014, from Digital Fordism Chapters: http://www.cddc.vt.edu/digitalfordism/fordism_materials/wsouza.htm Spindel, C. R. (1987). FormaÃ§Ã£o de um Novo Proletariado: As operarias do Distrito Ind ustrial de Manaus. SÃ£o Paulo: IDRC/FAPESP/IDESP. Suframa. (2010). Indicadores, de Desempenho do Polo Industrial de Manaus 1988 2010. Manaus: SuperintendÃªncia da Zona Franca de Manaus. Retrieved March 10, 2013, from suframa.gov.br: http://www.suframa.gov.br /zfm_projeto_oque.cfm Suframa. (2014). Indicadores de Desempenho do PÃ³lo Industrial de Manaus 2008 2014. Manaus: SuperintendÃªncia da Zona Franca de Manaus. Retrieved March 10, 2013, from http://www.suframa.gov.br/zfm_projeto_oque.cfm Suframa/Desenvovliment o_regional . (2014). Retrieved March 23, 2014, from Desenvolvimento Regional: https://www.suframa.gov.br/zfm_desenvolvimento_regional.cfm Suframa/HistÃ³ria . (2014). Retrieved March 12, 2014, from SuperintendÃªncia da Zona Franca de Manaus: suframa.gov.br/zfrm _historia.cfm The Economist. (2013, Sep. 28). Has Brazil blown it? The Economist . Retrieved from http://www.economist.com/news/leaders/21586833 stagnant economy bloated state and mass protests mean dilma rousseff must change course has Torres, I. C. (2005) . As Novas AmazÃ´nidas. Manaus: Editora da Univesidade Federal do Amazonas. Valle, I. M. (2007). GlobalizaÃ§Ã£o e reestruturaÃ§Ã£o produtiva: um estudo sobre a produÃ§Ã£o offshore em Manaus. Manaus: Editora Valera. Valle, Izabel M. (2007). GlobalizaÃ§Ã£o e Reestrut uraÃ§Ã£o Produtiva: um estudo sobre a produÃ§Ã£o offshore em Manaus. Manaus: Editora Valera. Veja. (2012, Nov. 23). ParticipaÃ§Ã£o do Sudeste no PIB encolhe, mostra IBGE. Veja: Economia .
19 7 Vidal, M. (2008). Taylorism. In V. N. Parrillo, Encyclopedia of Social Prob lems. Thousand Oaks, CA: Sage Publications. Volscho, T. W. (2008). Gini Coefficient. In Encyclopedia of Social Problems (pp. 408 409). Thousand Oaks: Sage Publications. Wesche, R. &. (1990). Integration and Change in Brazil's Middle Amazon. Ottawa: Univers ity of Ottawa Press. World Bank. (2012). World Development Report 2012: Gender Equality and Development. The Wolrd Bank. Washington DC: The World Bank.
198 BIOGRAPHICAL SKETCH Skyler Simnitt earned his Master of Arts in Latin American Studies from the University of Florida in 2014, and his Bachelor of Arts in Latin American Studies from Brigham Young University Provo, UT, in 2009. During his time as a Master of Arts student in the Center for Latin Amer ican Studies he was awarded the FLAS (Foreign Language and Area Studies) fellowship for the study of Haitian Creole. He also received the Tinker Travel Grant from the Center for Latin American Studies at the University of Florida in conjunction with his re search Master of Arts student he worked for the UF Center for Latin American Studies as an outreach associate, and research assistant for faculty.