Trading Goods or Human Capital

Trading Goods or Human Capital The Winners and Losers from Economic Integration Micha l Burzyński, Université catholique de Louvain, IRES Poznań University of Economics, KEM michal.burzynski@uclouvain.be KEM seminar Poznań, October 9, 2014

The outline of the talk: 1 Introduction 2 The Model 3 Calibration 4 Results 5 Conclusions

Migration in global demographics The share of immigrants in population Source: Trends in International Migration, UN

Migration in global demographics The number of immigrants from north/south Source: Trends in International Migration, UN

Stylized Facts and Motivation Increasing immigration is considered detrimental for the destination economies, whereas emigration is often disregarded. Although, the legal barriers for migration between many OECD countries are huge. Liberalization of trade is generally perceived as welfare enhancing. Consequently, tariff barriers for trade between the OECD countries are rather small, however non-tariff barriers are non-negligible. Both liberalization of migration and trade between developed countries are intensively discussed nowadays. For example: the Transatlantic Trade and Investment Partnership (EU and US) and the EU - TUR negotiations.

Motivation

The Research Questions The question: What are the economic consequences for the natives in OECD of liberalizing both migration and trade in the OECD?

The Research Questions The question: What are the economic consequences for the natives in OECD of liberalizing both migration and trade in the OECD? 1 Winners/losers of liberalization: comparison between countries and between education levels. 2 The consequences of local agreements: EU-US and EU-TUR. 3 The relations between migration and trade: (substitutability versus complementarity).

The Previous Estimations of Gains from Liberalization Global gains from liberalizing trade: estimated at the level of: 0.3% 5.2%, i.e. Anderson et al. 2000, 2006, Bouet 2008, Broda, Weinstein 2006, Francois et al. 2005, Hertel, Keeney 2005, Lai, Trefler 2004. Trade liberalization between EU-US: CEPR: 1.5% 2%. Ecorys: 0.3% 0.5%. Migration liberalization (world): extremely optimistic welfare gains: 67% Iregui 2005 to 122% Klein, Ventura (2007) or 147% Hamilton, Whalley 1984. Docquier, Machado, Sekkat 2012, Docquier, Delogu, Machado 2013: 2% 10% in terms of real GDP. di Giovanni, Levchenko and Ortega: the welfare impact of migration is 5% 10%.

Key Results in a Nutshell The main findings: Total gains from liberalizing migration: 2.0%, but: huge between-country disproportions in distribution. Total gains from liberalizing trade: 1.5%, but: gains are larger for the less integrated economies. In the EU-US and EU-TUR cases: the sending countries encounter losses (up to 6%), the receiving countries gain noticeably (almost 2.5%).

The Layout of the Model A multi-country, general equilibrium model assuming: homogeneous firms, heterogeneous labor, endogenous trade (exogenous shocks on bilateral trade costs), endogenous migration (exogenous shocks on bilateral migration costs). The Value Added: Liberalize migration with endogenous trade. Liberalize trade with endogenous migration. Concentrate on the currently discussed local agreements. Analyze the migration vs trade relations in a GE framework.

Assumptions Environment N countries indexed by i {1, 2,..., N}. Total, efficient labor endowment in country i: L T i. Labor is the only input for production.

Assumptions Environment N countries indexed by i {1, 2,..., N}. Total, efficient labor endowment in country i: L T i. Labor is the only input for production. Individuals Inelastically supply one unit of labor. Gain utility from consuming different varieties of goods. Are of type: native low-skilled, native high-skilled, foreign low-skilled, foreign high-skilled.

Individuals Preferences of agent of type s in country i are represented by a CES utility function over all the varieties available in country i (Dixit-Stiglitz type). The expenditures are equal to the nominal wage. The value of the indirect utility function: (equivalent to the real wage equivalent to the welfare) ui s = w i s. (1) P i Price index: N P i = B j (τ ij p j ) 1 ɛ j=1 1 1 ɛ.

Endogenous Migration Agents reach the decision to emigrate by comparing the real wages net the bilateral migration costs.

Endogenous Migration Agents reach the decision to emigrate by comparing the real wages net the bilateral migration costs. The agent s utility is a sum of two components: the deterministic one and the stochastic one (RUM), which results in heterogeneous preferences towards migration. U t ij = α ln [ ] w t ij ( ) 1 c t P ij + ε ij, t {l, h}. (2) i Migration cost: c t ij ( ; 1) and: j ct jj = 0. ε ij Gumbel(0;µ) (Extreme Value Type I Distribution).

Endogenous Migration Applying the McFadden s Theorem (1984) we get that: the probability of emigration of a person of skill t born in country j to country i is equal: [ πij t = Pr Uij t ( ) ] = max U t kj = k N ( ) exp Uij t ( N k=1 exp Ukj t ). (3) Labeling the number of people of skill t born in j, who live in country i by Mij t, we obtain: ( Mij t w t Mjj t = i /P i ( ) ) α/µ 1 c t wj t/p ij. (4) j

Firms In country i, there is a mass B i of homogeneous firms producing differentiated varieties of the consumption good. The firms production function is linear in labor: y i (k) = A i lt i (k). (5)

Firms In country i, there is a mass B i of homogeneous firms producing differentiated varieties of the consumption good. The firms production function is linear in labor: y i (k) = A i lt i (k). (5) The efficient labor composite is a nested CES with an imperfect substitution between low/high skilled as well as domestic/foreign workers: l T i (k) = ( θ H i ( l h i (k)) σ S 1 σ S + ( 1 θ H i ) ( l l i(k)) σ S 1 σ S ) σs σ S 1.

Firms Firms operate on a monopolistically competitive market, minimize the variable cost of production in their choice of optimal labor demand and maximize operational profits by setting the price: p i (k) = p i = ɛ ɛ 1 W i A i. However, there are barriers to entry, so that the profit is equal to zero, which leads to the market size equation: B i = L T i ɛf i. (6) In the equilibrium, the value of bilateral flow of goods is determined by the gravity equation: X ij X j = X i (P i /τ ij ) ɛ 1 N h=1 X. (7) ɛ 1 h (P h /τ hj )

Exogenous Variables The data describing the bilateral stocks of migrants are taken from Artuc, Docquier, Özden & Parsons (2014).

Exogenous Variables The data describing the bilateral stocks of migrants are taken from Artuc, Docquier, Özden & Parsons (2014). The data on GDP are from the World Bank Database.

Exogenous Variables The data describing the bilateral stocks of migrants are taken from Artuc, Docquier, Özden & Parsons (2014). The data on GDP are from the World Bank Database. Fixed cost: an unweighted synthetic indicator of three variables from Doing Business by the World Bank. the number of days needed to start business, the cost of starting a business (as a share of GNP p.c.), the survival rate of businesses, and the indicator is normalized, so that its minimum is 1.

Exogenous Variables and Parameters Bilateral trade data originate from CEPII.

Exogenous Variables and Parameters Bilateral trade data originate from CEPII. Trade and migration costs are calibrated numerically to fit the actual trade and migration data.

Exogenous Variables and Parameters Bilateral trade data originate from CEPII. Trade and migration costs are calibrated numerically to fit the actual trade and migration data. The TFP residuals are fitted numerically to obtain the general equilibrium in the system of N economies. Table 1: The Parameterization Description Symbol Value Elasticity of substitution between varieties ɛ 4 Elasticity of substitution between HS and LS σ 1.75 Elasticity of substitution between N and F σ M 20 Elasticity of agent s utility wrt real wage α/µ 1

The Liberalization of Migration Let us assume that the liberalization of migration takes place by an exogenous decrease in migration cost: cij t, for t {l, h}. However, this decrease concerns not the total cost, but its share which is due to the formal migration barriers (i.e. visa costs). An important issue is the identification of these costs. One can think of at least two strategies: using the survey data on intended migration (Gallup Survey), as in Docquier et al. (2013), a decomposition of the total migration cost using structural gravity regressions run on constructed (model) data.

The Liberalization of Migration Regression (for the low-skilled): ( ) ( ) M l ij w l i /P i ln Mjj l = ˆβ 1 ln wj l /P + ˆβ 2 ShortVISA ij + ˆβ 3 LongVISA ij + controls + FE (8) j

The Liberalization of Migration Regression (for the low-skilled): ( ) ( ) M l ij w l i /P i ln Mjj l = ˆβ 1 ln wj l /P + ˆβ 2 ShortVISA ij + ˆβ 3 LongVISA ij + controls + FE (8) j Table 2: Estimation of formal bilateral migration cost for the low-skilled Dependent variable: ( ) ln Mij l /Ml jj ( ) (1) (2) (3) (4) (5) ln w i l /P i w j l /P j 0.507 0.757 1.012 0.716 0.942 ShortVISA 0.553 0.261 0.531 0.500 LongVISA 0.976 0.467 0.355 0.607 Log distance 0.728 0.898 0.252 Log Pop ex 0.023 0.119 Log Pop im 0.807 Border 2.299 Language 1.565 Constant 8.415 7.517 2.247 13.623 4.344 exp FE Yes Yes Yes Yes Yes imp FE Yes Yes Yes Yes Yes Observations 1 190 1 190 1 190 1 190 1 190 Adjusted R 2 0.531 0.582 0.581 0.670 0.627

The Liberalization of Migration Regression (for the high-skilled): ( ) ( ) M h ij w h i /P i ln Mjj h = ˆβ 1 ln wj h/p + ˆβ 2 ShortVISA ij + ˆβ 3 LongVISA ij + controls + FE (9) j Table 3: Estimation of formal bilateral migration cost for the high-skilled Dependent variable: ( ) ln Mij h /Mh jj ( ) (1) (2) (3) (4) (5) ln w i h /P i w j h/p j 0.500 0.532 1.006 0.546 1.100 ShortVISA 0.696 0.361 0.478 0.473 LongVISA 0.788 0.495 0.377 0.708 Log distance 0.453 0.511 0.166 Log Pop ex 0.085 0.093 Log Pop im 1.090 Border 1.325 Language 1.546 Constant 7.747 7.283 4.351 22.674 4.911 exp FE Yes Yes Yes Yes Yes imp FE Yes Yes Yes Yes Yes Observations 1 190 1 190 1 190 1 190 1 190 Adjusted R 2 0.417 0.543 0.644 0.651 0.595

The Liberalization of Trade Regression 1 : ln τ ij = ˆβ 1 Barriers ij + β 2 ln (dist ij ) + controls + FE (10)

The Liberalization of Trade Regression 1 : ln τ ij = ˆβ 1 Barriers ij + β 2 ln (dist ij ) + controls + FE (10) Table 4: Estimation of formal bilateral trade cost Dependent variable: ln τ ij (1) (2) (3) (4) (5) Tariffs and NTB 1.718 0.343 0.703 0.600 0.629 Log distance 0.340 0.288 0.276 0.245 Log Pop ex 0.124 0.095 0.077 PPP im / PPP ex 0.299 0.288 0.264 Border 0.159 0.160 Language 0.273 0.230 Currency 0.144 Constant 1.546 0.926 1.103 0.810 0.751 exp FE Yes Yes Yes Yes Yes imp FE Yes Yes Yes Yes Yes Observations 1 190 1 190 1 190 1 190 1 190 Adjusted R 2 0.509 0.518 0.506 0.501 0.528

Liberalizing Migration - Aggregated Results Table 5: Demographic consequences of liberalizing migration EU OECD L l i L h i L l i L h i L l i L h i L l i L h i MIN -1.59% -4.02% 12.91% 14.99% -1.72% -2.24% 21.99% 20.89% MID -2.25% -6.20% 18.45% 23.28% -2.44% -3.47% 31.32% 32.33% MAX -4.19% -7.32% 34.88% 27.07% -4.46% -4.16% 57.39% 39.25%

The Demographic Impact of Migration - LS versus HS Figure 1: The demographic impact of liberalizing migration. Only a few countries experience a net inflow of workers (mainly from OECD).

Liberalizing Migration - Aggregated Results Table 6: Macroeconomic consequences of liberalizing migration EU OECD Real GDP Import Export Real GDP Import Export MIN -1.85% -1.32% -1.32% 1.40% -0.31% -0.31% MID -2.79% -1.93% -1.92% 2.01% -0.45% -0.45% MAX -3.57% -2.34% -2.33% 3.28% -0.28% -0.28%

The Welfare Impact of Migration - LS versus HS Figure 2: The welfare impact of liberalizing migration. The majority of countries lose, while there are only several winners, the HS workers are mainly better off.

The Net Flows of HS Workers after Liberalization Figure 3: The net flows of high-skilled workers.

The Net Flows of LS Workers after Liberalization Figure 4: The net flows of low-skilled workers.

Liberalizing Trade - Aggregated Results Table 7: Demographic consequences of liberalizing trade EU OECD L l i L h i L l i L h i L l i L h i L l i L h i MIN -0.01% -0.03% 0.29% 0.19% 0.04% 0.02% 0.06% 0.48% MID -0.02% -0.07% 0.53% 0.35% 0.07% 0.03% 0.09% 0.90% MAX -0.03% -0.08% 0.64% 0.42% 0.09% 0.04% 0.10% 1.08%

The Demographic Impact of Trade - LS versus HS Figure 5: The demographic impact of liberalizing migration. The majority of countries gets new immigrants (mainly from ROW).

Liberalizing Trade - Aggregated Results Table 8: Macroeconomic consequences of liberalizing trade EU OECD Real GDP Import Export Real GDP Import Export MIN 0.52% 3.95% 4.19% 0.79% 12.07% 12.07% MID 0.97% 7.29% 7.72% 1.51% 22.48% 22.48% MAX 1.17% 8.71% 9.23% 1.83% 26.99% 26.99%

The Welfare Impact of Trade - LS versus HS Figure 6: The welfare impact of liberalizing trade. All workers gain equally, LS are minimally better off.

Welfare gains from local agreements Table 9: Changes in real GDP changes after a full integration EU-US Trade Migration Trade and Migration EU US OECD EU US OECD EU US OECD MIN 0.22% 0.19% 0.15% -1.08% 1.09% 0.30% -0.86% 1.27% 0.45% MID 0.41% 0.35% 0.28% -1.67% 1.64% 0.44% -1.26% 1.99% 0.72% MAX 0.50% 0.43% 0.34% -2.19% 2.38% 0.70% -1.69% 2.80% 1.04%

The Welfare Impact of Integration between EU and US Figure 7: The welfare impact of liberalizing trade and migration.

Welfare gains from local agreements Table 10: Changes in real GDP changes after a full integration EU-TUR Trade Migration Trade and Migration EU TUR OECD EU TUR OECD EU TUR OECD MIN 0.02% 0.72% 0.01% 0.32% -2.40% 0.05% 0.34% -1.65% 0.06% MID 0.03% 1.36% 0.02% 0.47% -3.37% 0.08% 0.49% -1.93% 0.10% MAX 0.04% 1.64% 0.03% 0.82% -5.71% 0.14% 0.84% -3.92% 0.16%

The Welfare Impact of Integration between EU and TUR Figure 8: The welfare impact of liberalizing trade and migration.

The Relations between Trade and Migration Regressions: Trade ij(ji) = β 0 + β 1 Migration ij (11) where: Trade ij is the change in the share of trade from country j to country i to the total GDP in country i in percentage points, Migration ij is the change in the share of immigrants from country j in country i to the population in country i, in percentage points. Table 11: The Relations between Trade and Migration Migration liberalization Trade liberalization Dependent variable: Trade (i to j) Trade (j to i) Trade (i to j) Trade (j to i) Low-skilled (i to j) 0.08 0.06 15.88 11.57 High-skilled (i to j) 0.03 0.02 10.91 8.15

Conclusions The problem with liberalizing migration is the unequal distribution of the welfare gains across the OECD countries. People in the majority of countries (21 out of 34) are losing. The high-skilled are relatively better off, whereas the low-skilled lose drastically. Liberalizing trade does not bring large benefits, but it reduces the between-country inequality. Bilateral liberalizations (EU-US and EU-TUR) are positive for the destinations and harmful for the sources. Liberalizing migration spurs trade (complementarity), whereas liberalizing trade reduces migration (substitutability).

Thank you for your attention