Institutions in Context: Inequality Development, Politics, and Inequality in Latin America and East Asia Inyoung Cho DPhil student Department of Politics and International Relations University of Oxford
Latin America and East Asia Considerable socio-political similarities in both regions : History of rapid growth and industrialization as well as a long period of authoritarian military rule Considerable differences in regards to distributional results : Latin America: The most unequal region in the world : East Asia : Relatively equal region despite the long history of non-democracy or authoritarian rule 2
Puzzles Puzzle 1. Latin American countries have formed the most advanced welfare systems among the group of developing countries since WWII. However, welfare spending in Latin America did not contribute to the decline of inequality, especially during industrialization period. Why? Puzzle 2. Despite relatively abundant left power resources, why have Latin American countries become more unequal than their East Asian counterparts? Why did left power resources not lead to a more equal distribution of wealth? 3
Average Gini Coefficient during 1970s to 2000s 1960s 1970s 1980s 1990s 2000s Average Argentina 37.96 39.57 41.35 44.27 45.47 42.18 Brazil 40.19 57.40 51.06 51.76 49.41 51.51 Chile 48.35 43.38 51.40 51.72 50.03 49.65 Colombia 56.98 49.89 49.34 51.13 50.63 51.56 Costa Rica 42.46 41.41 41.97 42.63 45.23 42.87 Mexico 51.71 49.85 45.96 48.65 46.64 48.20 Peru 51.84 52.53 47.88 50.04 51.32 50.01 Average 44.76 46.24 45.80 47.04 47.05 46.55 Hong Kong 39.57 33.48 33.62 37.48 43.03 36.48 Indonesia 38.62 37.61 33.88 34.55 35.72 35.66 Japan 31.08 27.65 26.50 29.39 33.38 29.18 Korea 31.39 35.99 34.72 30.02 31.59 32.85 Malaysia 46.15 45.20 44.04 42.19 37.23 43.13 Singapore 32.57 30.84 33.28 34.88 38.39 34.06 Taiwan 33.58 26.61 26.84 28.07 30.84 28.35 Average 38.64 36.51 35.31 36.72 37.30 36.69 4
Research Question Q. Why are the levels of income inequality so different in Latin America and East Asia? Q. What can explain the different patterns of inequality in Latin American and East Asia? 5
Argument: Industrialization and Institutions Argument 1: Distinctive industrialization policies make a different impact on inequality. : Two dissimilar industrialization strategies, the ISI (Import substitution industrialization) and the EOI (Export oriented industrialization), explain the different levels of inequality. 6
The Development Strategies: ISI and EOI ISI Pursuing domestic market growth through protective industrial policies. Protectionist policies that include appreciated exchange-rate, extensive use of tariffs and quotas, complex multiple-exchange-rate systems, direct subsidies for firms, and state-owned banks are the core instruments of ISI. Latin American countries pursed the ISI model from the 1910s to the 1980s, mostly until the Debt crisis in 1982. EOI Aiming at economic growth through higher export based on price competitiveness of manufacturing. East Asian countries began adopting the EOI model in the period between the 1930s and the 1960s and maintained these strategies until the 1990s 7
The Development Strategies: ISI and EOI ISI and EOI : A set of industrial policies/strategies for economic growth that served as core frameworks of economy. ISI (Import Substitution) EOI (Export-oriented) Closed Economy Capital-intensive and Technology-intensive Manufacturing Import substitution production for domestic market growth Open Economy Light manufacturing for mass production Export oriented production for growth 8
Argument: Industrialization and Institutions Argument 1: Distinctive industrialization policies make a different impact on inequality. : Two dissimilar industrialization strategies, the ISI (Import substitution industrialization) and the EOI (Export oriented industrialization), can explain the different levels of inequality. Then, how? Via institutions! Industrialization affects institutions. The two development models had a different impact on economic structures, labor market institutions, and welfare institutions. In turn, the institutional configurations shaped by development models have different impacts on inequality in both regions. 9
Mechanism Economic Structure Development Strategy (ISI or EOI) Labor Market Institutions Inequality Welfare Policies 10
Argument: Industrialization and Institutions Argument 2 More importantly, different types of industrialization models lead to a different arrangement of institutions. Different industrialization models change the characteristics and ability of institutions. Thus the impact of institutions on inequality in EA and LA would be different. The choice of development strategy leads to a distinctive development of economic structures/ labor market institutions/welfare policies in Latin America and East Asia, as a result, it is connected to different level of inequality. 11
ISI/EOI and different institutional arrangements Economic Structure Labour Market Institutions Welfare Spending Pattern Relatively larger public and informal sectors (ISI) More urban and rural dichotomies (ISI) Relatively higher inflation (ISI) Higher insideroutsider bifurcation in the labor market (ISI) Less labor unions and lower union mobilization (EOI) Relatively Higher Spending on Education and Health (EOI) Higher Spending on Social Spending (ISI) Inequality Inequality Inequality 12
Mechanism: 1. Economic Structure Firstly, given its heavy emphasis on domestic market growth, the ISI model tends to develop larger public and informal sectors than the EOI model. : Public owned-firms and banks were established to support domestic industrial sectors. : Control of export and export-led sectors not only created inefficient distribution of the resources amongst sectors, but also the bigger size of informal economy, corruptions, and smuggling (Haggard, 1990). The bigger public and informal sectors lead to the unequal and inefficient distribution. 13
Mechanism: 1. Economic Structure Secondly, ISI tends to be more towards urban-rural dichotomy than the EOI model. Its capital-intensive production system required only skilled and semi-skilled labors, as a result, inevitably hindered the absorption of labor surplus from rural areas into the formal sector. The slowly increasing agricultural productivity and government subsidies mainly for producers and capital-intensive production raised urbanrural conflicts (Haggard, 1990) 14
Mechanism: 1. Economic Structure Lastly, ISI is associated with relatively higher inflation than EOI. Government and firms in Latin America were relatively tolerant to inflation and wage increases, because they were byproducts of the domestic market expansion. To the contrary, in East Asia, government controlled inflation harshly in order to maintain the price competitiveness in the export market. Successful control of inflation led to an increase of export, which gave them the capacity to absorb labor surplus especially from rural areas for more production and export (Fields 1994). 15
Mechanism 2. Labor Market Institutions Firstly, ISI tends to foster an insider-outsider bifurcation in the labor market. : The emphasis on capital-intensive and technology-intensive production demanded only skilled and semi-skilled labors, which consequently led to a downfall of traditional labor absorbing industries, and, as a result, to the dichotomy of labor market. : Also, the organized interests of urban formal sector workers tried to enact various protectionist policies in order to keep their interest. 16
Labor Market Institutions and Inequality Secondly, the EOI model is less favorable to labor unions and union mobilization than the ISI model. As a result of small-scale light manufacturing, the East Asian export-oriented economies required price competitiveness of labor costs and goods, and governments tried to control labor power and unionization. 17
A Choice of development strategy and Welfare Spending ISI is related to higher social insurance spending and EOI is associated with higher education and health spending. (Wibbels and Ahlquist 2007) Government developed different types of welfare state; which is more helpful to support the type of development a country adopt The capital and technology intensive production under ISI develops an insurance-based welfare scheme as a way to protect skill-owned labor forces. The labor-intensive and export-centric EOI model, on the other hand, develops universal primary education and healthbased spending scheme in order to supply and substitute proper manpower easily and to provide workers with minimum medical treatment, when they suffer from injuries obtained at the workplace. 18
Argument + The Impact of politics on Inequality To clarify the influence of political variables in inequality, three political variables are examined 1. Democracy 2. Strength of democratic tradition : Cumulated democracy score for the last 20 years from the observation year 3. Partisanship. 19
Mechanism: Politics via institutions Development Strategy (ISI or EOI) Economic Structure Labor Market Institutions Inequality Politics Welfare Policies 20
Measuring ISI/EOI ISI Numerical measurement: : The amount of manufacturing products consumed in domestic market, followed by Wibbels and Ahlquist (2007) : Higher number signifies higher tendency of ISI : The total manufacturing production is calculated as the sum of total value added based on the data provided by World Bank. 21
Summary of expected relationship The impact of institutions on Inequality Variables Latin America (ISI) East Asia (EOI) Dependent Variable Inequality Economic structure Inflation + + Urban-rural dualism + 0 Public employment size + - Informal sector size + +/0 Labor market institutions Union density +/0 -/0 Job protections for full time workers + +/0 Welfare spending Education and Health spending 0/- 0/- Social Security 0/+ - 22
Data and Analysis Empirical test: Hong Kong, Japan, South Korea, Philippines, Singapore, Thailand, Malaysia, and Indonesia in East Asia, and Argentina, Brazil, Chile, Costa Rica, Mexico, Uruguay, and Venezuela in Latin America. A unbalanced pooled time cross-sectional analysis since 1960s 23
Data and Analysis Two parts of analysis Firstly, test the relationship between development strategy and various institutions using Fixed effect generalized least squares (feasible FE GLS) after diagnostic tests Secondly, test the influence of those aforementioned institutions on inequality, using feasible Fixed effect generalized least squares (feasible FE GLS) with heteroskedasticity after diagnostic tests (Nunziata 2005) Diagnostic tests to check for auto correlations, contemporaneous correlation and heteroskedasticity: used a series of the Breusch and Pagan s Lagrange Multiplier test, Hausman test, modified Wald test. From this part of the analysis, I indirectly see the impact of development strategy on inequality. 24
Data and Analysis 25
Data and Analysis : Methodology First regression model: Second regression model: All variables excepting dummy variables are logarithmic transformed. 26
Result: FGLS Regression 1 27
Result: FGLS Regression 2 Dependent Variable Independent Variables Inequality (Gini) Coef. P> z Inflation 0.071 (0.014) 0.000 Public employment 0.006 (0.015) 0.714 Urban-rural dualism 0.224 (0.0732) 0.002 Union density -0.176 (0.0305) 0.000 Protections for full time workers 0.197 (0.0612) 0.001 Education and Health spending -0.036 (0.0324) 0.265 Social Security Spending 0.014 (0.020) 0.485 GDP per Capita -0.007 (0.0277) 0.806 GDP growth 0.089 (0.034) 0.009 Unemployment 0.147 (0.026) 0.000 Female labor force participation -0.304 (0.046) 0.000 _cons 5.180 (0.431) 0.000 28
Result: FGLS Regression 3 Asia Latin Dependent Variable Inequality (Gini Coefficient) Independent Variable Inflation -0.007 (0.044) 0.014 (0.012) Public employment -0.004 (0.028) -0.037*** (0.012) Urban-rural dualism -0.527*** (0.108) 0.147*** (0.041) Union density -0.096*** (0.028) -0.011 (0.026) Labor market regulations 0.258*** (0.074) 0.101** (0.051) Education and Health spending -0.072*** (0.020) -0.059*** (0.023) Social Security Spending -0.042** (0.021) 0.051*** (0.011) GDP per Capita -0.192*** (0.040) 0.019 (0.052) GDP growth 0.033 (0.026) 0.020 (0.025) Unemployment 0.070*** (0.031) -0.021 (0.014) Female labor force participation 0.507*** (0.093) -0.080** (0.036) 29
Result: F-Test for Equality of coefficients Variables P-values Inflation 0.7454 Public employment 0.6124 Urban-rural dualism 0.0000 Union density 0.0037 Labor market regulations 0.0086 Education and Health spending 0.0000 Social Security Spending 0.0023 GDP per Capita 0.0000 GDP growth 0.4236 Unemployment 0.0102 Female labor force participation 0.0000 30
The Direct impact of politics on inequality Independent Variables Asia Latin Democracy 2.95 (0.87)*** 0.24 (0.13) Democratic History -0.33 (0.06)*** -0.14 (0.02)*** Partisanship (left to the right) 0.02 (0.04) 0.03 (0.01)*** Inflation (t-1) -0.04 (0.05) -0.01 (0.01) Public employment 0.05 (0.04) -0.04 (0.01)*** Urban-rural dualism -1.13 (0.33)*** 0.02 (0.04) Union density -0.20 (0.04)*** -0.01 (0.02) Protections for full time workrers 0.12 (0.10) 0.00 (0.04) Education and Health spending 0.01 (0.04) 0.01 (0.03) Social Security Spending -0.20 (0.07)*** 0.01 (0.02) GDP per Capita -0.51 (0.13)*** -0.12 (0.05)*** GDP growth 0.09 (0.03)*** 0.00 (0.02) Unemployment 0.05 (0.04) -0.01 (0.02) Female labor force participation 0.61 (0.15)*** 0.08 (0.04)** _cons -1.18 (1.37) 4.61(0.47) 31
Findings Empirical results partially support the argument linking development model, different institutional arrangements, and inequality. All in all, the development model hypothesis well explains insider-outsider inequality through labor market institutions. The Development model hypothesis is not very successful in linking development strategies to inequality via economic structure. 32
Conclusion it clearly shows that developmental model is related to several institutional variables that are known to be important determinants of inequality. Although not all of my theoretical expectations on the relationship between development strategy and institutions coincide with the empirical testing results, the second FGLS result shows that the determinants of inequality in Latin America and East Asia are quite different, and the directions and impacts of the determinants on inequality are also entirely dissimilar between Latin America and Asia. ISI and EOI have different influences on the same institutions and affect the ability and characteristics of the institutions differently. 33
Thank you! 34