Income Inequality, Trade and Financial Openness

Income Inequality, Trade and Financial Openness G.C. Lim and Paul D. McNelis January 214 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 1 / 34

Order of Presentation Introduction Empirical analysis Model: data set of low to middle income countries (based on GDP per capita) panel data results DSGE model with income distribution small open economy with natural resources labour-rich, capital poor Simulation analysis insights into relationships between openness and Gini role of labour intensities Concluding remarks G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 2 / 34

Income Inequality: Changing Tolerance Hirschman and Rothschild (1974): diminishing tolerance for the tunnel effect Kuznets: inverted U curve between inequality and growth. Dual economy hypothesis Kaldor: inequality helps development because rich save more Deaton and the Great Escape: some are left behind Dollar and Kray: growth is good for the poor Question: are trade and financial openness more effective for reducing inequality? Openness defines with respect to trade, foreign aid and FDI. G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 3 / 34

Globalization Jaumotte, Lall and Papageorgiou (28): inequality Globalization has ambiguous effect. technology increases Trade reduces inequality while foreign investment increases inequality If openness does not matter, education and technology are the remaining options Which mechanisms ensure that growth (through trade or financial liberalization) reduce inequality? G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 4 / 34

Data Sets Poorest is Tanzania with 1% of world s mean per-capital GDP. Brazil has highest Gini, Bulgaria has lowest Brazil has the lowest index of openness (2.5%), Honduras the highest (12.3%) Bangladesh has virtually no foreign investment Foreign aid for Mozambique is more than 33% of GDP G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 5 / 34

Statistical Properties of Indices (Percentage) Gini dy ry gy op ay fy Full sample (42) mean 43.323 1.375 67.616 17.237 65.76 5.234 2.763 median 43.39 1.423 66.178 14.923 59.253 1.652 2.55 Low-Income Countries (12) mean 42.485 1.31 22.386 17.736 56.672 13.162 2.777 median 42.93.957 22.378 15.921 55.475 11.137 2.21 Lower-Middle Income Countries (2) mean 41.431 1.33 67.374 17.949 67.176 2.919 2.763 median 41.6 1.33 67.766 14.764 61.556 1.593 1.71 Upper-Middle Income Countries (1) mean 48.113 2.147 122.377 15.216 7.96.35 2.747 median 47.438 2.121 124.88 14.335 64.24.163 2.645 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 6 / 34

Correlation Coeffi cients with the Gini (Pooled Data Sets) dy ry gy op ay fy Full-set -.48.165* -.142* -.132*.48 -.15 Low -.28*.43.189* -.12.259*.185* Lower-middle.1 -.244* -.325* -.2.28* -.57 upper-middle -.299*.243*.5* -.726* -.378* -.141** Note: * Significant at the 5% level G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 7 / 34

Correlations with the Gini (summary of country-specific results) Openness FA FDI Full set of 42 countries mean.185 -.99.82 minimum -.911 -.887 -.74 maximum.958.894.868 Low Income mean.49.34.88 minimum -.911 -.887 -.53 maximum.958.894.677 Lower-Middle mean.474 -.119.88 minimum -.887 -.689 -.48 maximum.917.679.791 Upper-Middle mean -.229 -.218.62 minimum -.885 -.75 -.74 maximum.952.67.868 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 8 / 34

Panel estimation (with country and time specific fixed effects): GN it = α i + α t + β 1 dy it + β 2 ry it + β 3 gy it + β 4 op it + β 5 ay it + β 6 fy it + ɛ it ; ɛ it N(, σ 2 ɛ ) Fixed Effects Panel Estimates: 1992-27 dy it ry it gy it op it ay it fy it Full-dataset -.197* -.33**.29*.51*.37.75 Low-Income -.429*.354* -.63.9*.23.84 Lower-Middle -.155* -.2.23*.99* -.189* -.49 Upper-Middle.39 -.95* -.158 -.68* -1.176*.174** Note: Significant at 5% level; Significant at 1% level G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 9 / 34

Gini, Growth, Relative Income and Government Spending Key result of panel: effects on inequality growth and relative income have significant For relatively more prosperous, relative income more important For low-income countries, growth of income reduces inequality In these countries, widespread increase in income needed to reduce poverty and inequality. Higher government spending as a share of GDP, increases inequality (Rudra 24) She argued that spending programs in the lower income countries have much greater political lobbying There are thus little redistributive effects of public spending G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 1 / 34

Gini and FDI Positive effect of FDI on Gini Corroborates positive results of Feenstra and Hanson (1997) on wage inequality in Mexico Basu and Guarigliab (27) : FDI reduces agricultural share of GDP G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 11 / 34

Gini and AID Little evidence that aid reduces inequality, similar to Calderón and Chong (26) Deaton: aid which reduces mortality rates increases the population of poor people. Perverse effects: access to foreign aid flows makes governments less accountable to domestic constituents, G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 12 / 34

Gini and Trade Openness Trade openness has negligible effects for low income countries, but has significant negative effects for the middle income countries Gourdon, Maystre, and de Melo (28): initial endowments with respect to skilled labour, matter most. Acar and Dogruel (21) for MENA: openness reduces inequality Munschi (212) on Bangladesh: labour intensive industries Turnovsky (213): greater openness led to increase in many stories to tell based on empirical results Ehrlich and Kim (1997), these empirical results provide evidence of empirical associations, G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 13 / 34

DSGE Model: Key Features Stylised low-income economy Households heterogenous agents with different labour endowments Gorman polar form utility function Production of Two Types of Goods tradeables with prices determined globally non-tradeables with market-clearing prices Financial sector accepts deposits from households, borrows from foreigners lends to public sector and firms Public sector manages government expenditures taxes labour income and consumption sets exchange rate G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 14 / 34

Heterogeneous agents with one unit of time, for work and leisure Gorman Polar Form Utility: ( ) 1 max C,L,K,ME β t h t= η (C t) η (V t ) ωη G χη t (1) Consumption is a basket of traded and non-traded goods Households accumulate capital and rent to firms: K h t = (1 δ)k h t 1 + I h t (2) Investment is imported Parameter δ determines the replacement rate of imported capital goods G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 15 / 34

Budget Constraint and Euler Equations Household budget constraint: (1 τ w )W t (H V t ) + (1 + R m t 1)M t 1 + (3) R k t K h t + Π t = (1 + τ c )P c t C t + M t + P k t I t (4) Euler Equations: (C t ) η 1 (V t ) ωη G χη t (1 + τ c )P t = E t ωc t = (1 τw )W t V t (1 + τ c )Pt c [ ( ) [ ] (1 + Rt m ) Pt k Rt k = E t Pt+1(1 k δ) β(1 + R m t ) (C t+1) η 1 (V t+1 ) ωη G χη t+1 (1 + τ c )P c t+1 (5) ] (6) (7) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 16 / 34

Exports: natural resource products Y x t = C x t + X t ln(x t ) = ρ x ln(x t 1 ) + (1 ρ x ) ln(x ) + ɛ x t, ɛ x N(, σ x ) (9) Production function: Price of export goods: (8) Y x t = Z x (K x t ) αx (L x t ) 1 αx (1) ln(pt x ) = ρ p ln(pt 1) x + (1 ρ p ) ln(p x ) + ɛt p, ɛ p N(, σ p ) (11) Wage bill (financed by borrowing from banks): N t = W t L x t (12) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 17 / 34

) α Yt h = G k Z (K h t h h ( ) 1 α L h h t (13) Y h t = C h t + G t (14) G t = G c t + G k t (15) G k t = ςg (16) Government spending is for consumption and for infrastructure G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 18 / 34

Pricing, labour Mobility Flexible prices for home goods First order conditions: Π h t = P h Y h W t L h t R k t K t (α h )W t (1 α h )R k t = K h t L h t Labour is mobile across the two sectors Marginal costs: P h t = (W t) 1 αh (R k t ) αh Z h t G k t ( 1 (α h ) αh (1 α h ) 1 αh ) (17) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 19 / 34

Banks and Financial Sector Reserves and Lending Costs: Φ m t = Φ m + ϕ m (M t 1 M) (18) Φ n t = Φ n + ϕ n (N t 1 N) (19) Risk premium on foreign borrowing: Φ s t = Φ s + ϕ s (F t 1 F ) (2) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 2 / 34

Interest Rates and Macroeconomic Identities FOC s of banks: Foreign Debt (1 + Φ n t ) (1 + R t ) = (1 + R n t ) (21) (1 Φ m t ) (1 + R t ) = (1 + R m t ) (22) (1 + R t )S t = (1 + R t + Φ s t )S t+1 (23) SF t = (1 + Rt 1 + Φ s t 1)SF t 1 + SPt m I t SPt x X t (24) Domestic Debt: B t = (1 + R t 1 )B t 1 + P h t G t τ w t W t L t τ c t P t C t Q t P z t K x t (25) Q t = Φ m t M t + Φ n t N t + (1 + R n t )N t (1 + R n t 1)N t 1 (26) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 21 / 34

Openness, FDI, and Aid Openness and GDP FDI Modification Φ t = S tpt m Y t = P h t Y h t I t + S t Pt x X t Y t + P x t Y x t (27) foreigners own capital capital disappears from household budget constraint investment decision is discounted by return to portfolio investment AID Modification: foreigners pay for investment goods funds go to the government funds going to government sector G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 22 / 34

Initial Endowments Each agent receives a share of initial deposits (money endowments) and share of profits: Π i t = h i Π t (28) [ Mt i = (1 + Rt 1)M m t 1 i + (1 τt w )W t 1 (1 + ] ω)ρi V t [ ω ] + h i Π t + h i Rt k Kt h Pt k I t (29) (3) y i t = (1 τ w )W t (1 ρ i V t ) + (1 + R m t 1)M i t 1 + h i R k t K h t + h i Π t (31) Money endowments and leisure relation: ρ i = 1 V ω R m M i + (1 τ w )W + h i Π + h i [ R k P k δ ] K h (1 + ω) (1 τ w )W (32) G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 23 / 34

Lorenz Curve 1 Lorenz Curve cumulative share of income from lowest to highest 9 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 9 1 Cumulative Share of Income from Lowest to Highest Implication: richer members of the economy work less. Income of about 9 per cent of workers is through the provision of labour services. G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 24 / 34

Inequality and Steady States Sums of Labour, Income and Money equal Aggregate Steady States: Deaton modified Gini index: H Mt i = M (33) i=1 H Vt i = H L (34) i=1 DG = H + 1 H 1 2 H i=1 p i y i (H 1) H i=1 y i G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 25 / 34

Parameter Specification Parameters Definitions Calibrated Values β discount factor 1/1.4 η relative risk aversion -.5 ω labour supply elasticity.5 χ government spending in utility.15 γ share of tradeables in consumption bundle.5 θ intratemporal substitution elasticity 2.5 ϕ m, ϕ n, ϕ s risk premium parameters.1,.1, ρ x, ρ p autoregressive terms for shock processes.5,.5 σ x, σ p standard deviation for shocks to X, P x.1x,.1p x τ w, τ c tax rates.1,.5 ς share of public capital expenditure in G.12 δ share of import replacement in capital α h, α x parameters in production function.5,.5 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 26 / 34

Implications of Parameter Calibration Calibration is set so that the index of openness is.513, close to observed mean in the data Ratio of government spending to GDP is.31. Deaton-adjusted Gini is.49, within the range of the reported sample. Share of labour in non-traded sector is 65%, share of labour income is 5%, Consumption spent on traded goods is 1%, proportion of non-traded G is 39% Most traded goods sold overseas: 87% Arellano (29) parameterizes the non-traded sector as more capital intensive. Kuralbayeva and Vines (28) assume the reverse G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 27 / 34

Impulse Response: Export Demand Shock 1 Export Volume Shock.15 Consumption Price Index.8 Consumption 5.1.5.6.4.2 1 2 1 2 1 2.4 Wage Income 4 Money Balances.5 Profits.2 3 2.5 1.2 1 2 1 1 2 1.5 1 2 15 Output 15 x 1 3 Openness Index.1 Gini 1 5 1 5.5.5 1 2 5 1 2.1 1 2 G.C. Lim and Paul Impulse D. McNelis Reponses () Incomefollowing Inequality, Tradea and shock Financialto Openness Export Demand January 214 28 / 34

Impulse Response: Terms of Trade Shock: Export Price Shock Consumption Price Index.2.3.5 Consumption.15.1.5 1 2.2.1 1 2.5 1 1 2.4 Wage Income 4 Money Balances 1 Profits.2 3 2 1 1 2.2 1 2 1 2 3 1 2 15 Output.3 Openness Index.1 Gini 1 5.2.1.5.5 1 2.1 1 2.1 1 2 Impulse Reponses following a shock to Export Price G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 29 / 34

Impulse Response with and without FDI and Foreign Aid.1 Export Demand Shock.1 Export Price Shock.5.5.5 Base With FDI.1 5 1 15 2.5 Base With FDI.1 5 1 15 2 6 x 1 3 Export Demand Shock 4 2 2.1.5 Export Price Shock 4 6 Base With FA 8 5 1 15 2.5 Base With FA.1 5 1 15 2 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 3 / 34

Stochastic Simulation Experiments: Design We are interested in the relations between the Gini and openness Model re-parameterized for varying degrees of productive activity. Shocks processes remained the same Deep parameters fixed, as well as tax policy and risk premia parameters. Generate a number of stylised economies with varying degrees of openness We varied the factor share of capital in the production functions α h, α x, as well as δ These parameters affected the import content and degree of openness. G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 31 / 34

Sensitivity Analysis Sensitivity Analysis Average Correlation parameters case Gini Φ o t {Gini,Φ o t } δ αh α x 1.277.255.398.1.2.2 2.215.257.599.1.2.8 3.469.541.779.1.5.5 4.293.33 -.156.5.2.2 5.229.37 -.19.5.2.8 6.489.626 -.23.5.5.5 7.692.859.383.5.8.8 8.295.31 -.68.9.2.2 9.231.313 -.24.9.2.8 1.492.637 -.27.9.5.5 11.69.851 -.491.9.8.2 12.696.872 -.621.9.8.8 G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 32 / 34

Simulation Insights For economies with a high degree of openness, there is a negative correlation between the Gini and the index of openness, for high depreciation and high capital intensity in production For high openness and low capital intensity, and depreciation, there is a positive but small correlation between openness and the Gini. Despite high capital intensity in both sectors, the economy with the higher import content in the non-traded sector, opens the economy to more international shocks and the gains from trade are distributed more broadly. G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 33 / 34

Concluding Remarks Empirical results growth reduces inequality trade openness can increase or decrease the Gini FDI and AID have negligible effects for low-income countries but reduce inequality in middle-income countries Simulation results Key channel from open to Gini: distribution to labour income, with high capital intensities and high capital-replacement rates Under high openness and low capital intensity, and import-replacement rates, there is a positive but small correlation between openness and the Gini. G.C. Lim and Paul D. McNelis () Income Inequality, Trade and Financial Openness January 214 34 / 34