Policy Research Working Paper 8975 Can Taxes Help Ensure a Fair Globalization? François Langot Rossana Merola Samil Oh Poverty and Equity Global Practice August 2019 Policy Research Working Paper 8975 Abstract This paper analyzes whether taxation can be successfully disparities. In this context, the paper analyzes whether a used to reduce the incidence of labor informality and budget-neutral tax reform—switching the tax burden from achieve higher equality in a globalized economy. To this payroll taxes paid by firms operating in the formal sector to purpose, it develops a two-area model: a developed country a consumption tax—can mitigate possible adverse effects and an emerging country. The two areas differ according of trade liberalization and support labor formalization. The to the size of the informal sector, which is characterized effects of taxation are seen in the context of the trade-offs by a more flexible labor market and lower productivity. between growth, labor formality and equity. The analysis To illustrate the potential role of taxation in achieving a suggests that small improvements in formalization, result- more fair income distribution, the paper introduces a trade ing from the tax reform, come at the cost of widening shock to simulate the effects of trade liberalization. Trade income inequality. To reduce the incidence of low-quality expansion has often been blamed for leading to an expan- jobs, tax policy interventions should go hand in hand with sion of the informal sector and a widening of wage income more effective social protection systems and labor laws. This paper is a product of the Poverty and Equity Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at merola@ilo.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team ∗ Can taxes help ensure a fair globalization? François Langot† Rossana Merola‡ Samil Oh Ÿ ∗ Funding from the World Bank-ILO Research Fund is gratefully acknowledged. The work greatly ben- eted from helpful comments by Stéphane Adjemian, Thomas Brand, Matteo Cacciatore, Valerie Cerra, Ralph Chami, Olivier Charlot, Ekkehard Ernst, Aurelien Eyquem, Miguel Foguel, Fabio Ghironi, Jinill Kim, Stefan Kühn, Gladys Lopez-Acevedo, Hugo Ñopo, Pierella Paci, Ugo Panizza, Stela Rubinova, Raymond Saner, Thepthida Sopraseuth, Christian Viegelahn, Sergio Vieira, the DYNARE team, as well as partici- pants in seminars and conferences at the World Bank, the ILO, the Graduate Institute of International and Development Studies in Geneva, the International Labour and Employment Relations Association (ILERA) 2018 Congress, the 2018 Asian Meeting of Econometric Society, the 2019 Congress of the Swiss Society of Economics and Statistics and the EcoMod 2019 Conference. However, all errors remain our own. The views contained here are ours own and not necessarily those of the ILO and the World Bank Group. † Le Mans University (Gains-TEPP & IRA), Paris School of Economics & Cepremap. E-mail: angot@univ-lemans.fr ‡ ILO, International Labour Oce, Research Department. E-mail: merola@ilo.org Ÿ Bank of Korea, Research Department. E-mail: samil.oh@bok.or.kr Contents 1 Introduction 1 2 Model 3 2.1 Households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.2 Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.2.1 Intermediate goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2.2 Final goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.3 Government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.4 Closing conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3 Calibration 14 4 The impact of trade liberalization 17 5 Tax reform 22 6 Conclusions 27 A Dynare equations 34 B Steady states 37 2 List of Figures 1 The nal good sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2 The labor market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3 Unemployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 4 Wage inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5 Goods markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 6 Labor markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 7 Unemployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 List of Tables 1 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3 1 Introduction In the three decades prior to the global nancial crisis, household income inequality increased in a large majority of OECD countries. Emerging countries have income inequality signi- cantly higher than the OECD average, although some countries (such as Brazil, Indonesia and, on some indicators, Argentina) have recorded signicant progress in reducing inequality over the past 20 years (OECD, 2011). th In addition, during the late 20 century there was a general increase in the informal economy in many countries around the world (see Schneider and Enste (2000)). Heintz and Pollin (2005), for example, show that within a data set of 23 countries, 19 showed increases in informality. Similarly, ILO data show that from 2002 self-employment increased in all developing regions, and world-wide it increased from about one-quarter to one-third of non- agricultural employment during 1980-2000 (ILO, 2013). The high incidence of informality is an issue of concern especially in developing countries, where on average, more than 50 1 percent of the labor force is informal. Among the drivers of rising informality, the conventional view blames trade liberalization for being responsible for the fall in wages for unskilled and low-income workers, as well as the rise in informal and less protected forms of employment. Therefore, despite its uncontrover- sial expansionary eects on global growth, trade expansion has not always been translated 2 into more equal incomes and better working conditions. In these circumstances, an ecient tax system is indeed an important tool for addressing rising inequality and informality and restoring robust economic growth. On the one side, taxation is a powerful policy tool to redistribute income and make the post-tax income distribution less unequal. On the other side, taxation is a potential tool to lessen the costs of operating in the formal sector, since formality choices are very elastic to marginal tax rates. Indeed, a targeted and well-balanced tax code is an essential element in making further progress in achieving the Sustainable Development Goals by providing a stable funding base for high-quality public services for all and eective transfers targeted to those most in need. In this context, our analysis contributes to the literature by investigating whether tax- ation may be an ecient policy tool to support labor formalization in a globalized econ- omy without widening income disparities. Our analysis is based on a Dynamic Stochastic General Equilibrium (DSGE) model with two asymmetric countries: a developed and an emerging country. These two countries dier according to the size of the informal sector in the intermediate-good sector. The informal sector is characterized by a more exible labor market (i.e. rapid entry and exit and more exible adjustment to change in demand) and lower productivity. In this respect, our paper relates to the recent theoretical literature em- bedding the informal sector in DSGE models (e.g. Conesa et al. (2002), Busato and Chiarini (2004), Orsi et al. (2014), Pappa et al. (2015), Dellas et al. (2017)). Within this strand 1 In many Latin American countries informal employment exceeds 50 percent of total urban labor force (Gasparini and Tornarolli (2007)). Estimates for Sub-Saharan Africa and Asia are even higher (Jütting et al. (2008)). For an overview on job quality in emerging economies, see OECD (2015). 2 For a more comprehensive discussion, see Bacchetta et al. (2009). 1 of literature, very few works enrich DSGE models with both informality and a fully-edged labor market with search and matching frictions. The few exceptions, to the best of our knowledge, are Cook and Nosaka (2005), Zenou (2008), Satchi and Temple (2010), Batini et al. (2011), Colombo et al. (2018), Bosch and Esteban-Pretel (2015) for Mexico, Anand and Khera (2016) for India and Poirier and Trupkin (2018) for Argentina. However, most of the aforementioned theoretical works have a dierent focus and very often analyze the role of regulation, while none of them analyzes the eect of taxation on informality. In addition, models developed in the aforementioned studies describe closed economies and none of them is suitable to analyze the impact of policies in a globalized economy from both a developed and a developing country perspective. The novelty of our paper is to focus on the role of taxation in reallocating labor between formal and informal activities in countries which participate in international trade. The ef- fects of taxation are seen in the context of the trade-os between growth, labor formality and equity. We start from a model à la Melitz (2003) and we extend it in two directions. First, we propose a dynamic model as in Ghironi and Melitz (2005), with search and matching frictions as in Helpman, Itskhoki and Redding (2010) and Cacciatore and Ghironi (2015). Second, we distinguish two asymmetric areas, a developed and an emerging country, charac- terized by dierent incidence of informality. We model informality as proposed by Charlot, 3 Malherbet and Terra (2015) in a closed economy static model. Our model closely follows Cacciatore and Ghironi (2015). However, Cacciatore and Ghironi (2015) focus on developed economies where a representative agent can be employed in only one sector, the presence of informality being not considered. In order to fully capture the impact of a tax reform in emerging economies, it seems to be crucial to model the interplay between the formal and informal sectors. Therefore, our main contribution is that we embed the informal sector, as we believe that the analysis of labor market dynamics cannot be limited solely to the formal sector, given the high incidence of informality especially in developing and emerging countries. Furthermore, in order to assess whether a scal reform can enable transition to formalization, we add taxation as well as hand-to-mouth agents in the model, which is not embedded in Cacciatore and Ghironi (2015). Our work is related to the literature analyzing the impact of taxation on informality. Em- pirical evidence points out that reducing taxation on formal businesses eases the migration of entrepreneurs from the informal to the formal sector, where productivity is higher, with positive eect on output and economic eciency (see Slonimczyk (2012) for the Russian Fed- eration and Araujo and Rodrigues (2016) for Brazil). Higher tax rates among rm-owners induce not only substantial movements to the informal sector, but also under-reporting of taxable earnings and income shifting to tax-favored business forms, which may ultimately lead to inecient allocation or resources (see Waseem (2018) for an analysis of the Pakistani tax reform introduced in 2009). If informality is voluntary, lower taxation rates should reduce rms' incentives to enter the informal sector. However, even if informality is involuntary, lower tax rates could reduce informality by encouraging formal sector rms to expand em- 3 These authors limit their analysis to a closed economy static model. 2 ployment and create more formal jobs. This strand of the literature suggests that the best approach to reduce the size of the informal sector is using taxation to reduce the costs of being formal and create the right incentives for companies and workers intending to switch to the formal sector. However, dierent tax instruments may have dierent eects in pro- moting the transition to the formal economy. For instance, lower taxes on social security contributions or on capital translates to a lower degree of informality, whereas cutting taxes on labor income has the opposite eect. Reducing personal income taxes increases gross wages, thus making the informal sector  which is labor intensive  more attractive. Our analysis highlights a number of interesting results. We start by introducing a trade shock which simulates the eect of trade liberalization in a globalized economy. Simulations in our model point out that in the short term trade liberalization boosts economic activity and employment in both the formal and informal sectors. However, this employment ex- pansion is biased toward the informal sector, which is not subject to labor regulation and hence is more exible. In addition, in the long run  after the strong employment gains recorded during the initial phase of trade expansion  there is a phase characterized by a contraction on the labor market. We then investigate whether it is possible to correct this bias in favor of the informal sector by reducing payroll taxes paid by rms operating in the formal sector. This policy exercise simulates the eects of several programs implemented in emerging economies (e.g. SIMPLES and SUPERSIMPLES in Brazil and the Monotax in Argentina) aiming at reducing the tax burden for small enterprises which are more likely to 4 operate in the informal sector. We show that an increase of the consumption tax could be a relevant strategy to nance the payroll tax cuts. Although this budget-neutral tax reform supports the formalization process on the labor market, we observe that small improvements in formalization come at the cost of widening income inequality. The rest of the paper is structured as follows. In Section 2 we describe the model, while in Section 3 we discuss the calibration. In Section 4 we introduce the trade shock to simulate the eect of trade liberalization. The impact of a budget-neutral tax reform is discussed in Section 5. Finally, Section 6 concludes. Some technical aspects are reported in the Appendix. 2 Model We develop a two-country model, calibrated on a developed and an emerging country. The two economies are modeled exactly symmetrically, so that the following description in this Section holds for both economies. We assume a dual labor market, with formal and informal workers. The two countries dier for the incidence of labor informality, since emerging coun- tries are characterized by higher informality than advanced economies. Variables appearing with an asterisk refer to the modeled foreign economy. 4 These types of programs, such as microcredit and tax reliefs for small enterprises, have been blamed for limiting growth opportunities in emerging economies, since they increase the incentives for small enterprises to remain small (see Hsieh and Olken (2014)). For an analysis on the diculties of Brazilian small rms to surpass the threshold of medium-size plants, see Coelho et al. (2017). 3 There are four actors in each country: households, rms producing intermediate goods, rms producing nal goods and the government. The model features heterogeneous house- holds: Ricardian and non-Ricardian. Ricardian households hold bonds but do not supply labor, whereas non-Ricardian households do not have access to nancial markets to nance their consumption needs. Therefore, non-Ricardian households need to supply labor in order to nance their consumption needs. Since recent evidence shows that there is no segmen- tation between the formal and the informal sector (see Charlot et al. (2015)), we assume 5 that workers can move between the two sectors. They may decide to either supply labor in the formal sector, or supply labor in the informal sector or be unemployed. Labor is hence supplied only by non-Ricardian households to intermediate good producers. Inter- mediate good producers operate in a perfect competitive market and hire labor  either on the formal or informal market  to produce intermediate goods which are sold to nal good producers. Final good producers combine intermediate goods into a nal good which is sold 6 on a monopolistically-competitive market. Finally, to provide public goods and unemploy- ment benets, the government collects taxes paid on consumption by all households as well as payroll taxes paid only by employees and employers (i.e. intermediate good producers) operating in the formal sector. For the sake of simplicity, the model does not feature nominal price rigidities and goods are produced using only labor without capital. 2.1 Households There are two types of households in the economy: Ricardian and non-Ricardian. Ricardian households (indexed by a ) do not work, hold assets and have access to international nancial markets. Non-Ricardian households supply labor, but have no access to nancial markets. Non-Ricardian households can work in the formal sector (indexed by F ), work in the informal sector (indexed by I) or being unemployed (indexed by u). For all agents, the consumption basket Ct aggregates Home and Foreign consumption in a Dixit-Stiglitz form: φ 1 φ−1 φ−1 Ct = Ct (i) φ di (1) 0 5 For the sake of simplicity, we focus only on labor informality and we abstract from business informality, i.e. we abstract from modeling how rms can switch from the formal to the informal sector and vice versa. Chacaltana et al. (2018) show that business informality does not imply labor informality and vice versa. The decision of a rm to go formal is the product of a complex evaluation based not only on the tax burden but also on other factors, such as the opportunity to have access to credit. Modeling rms' choice to switch between the formal and informal sector would require embedding the nancial sector into the model, which will pose challenges to the analytical manageability of the model. Becker (2018) provides an example of a model featuring a sector-switching mechanism. 6 The distinction between formal and informal labor arises only for rms producing intermediate goods which are used by nal good producers as the sole input. Intermediate good producers are not allowed to directly export abroad. This assumption is needed because exporting means some minimal formality and respect of customs requirements and are more subject to control and customs inspection. 4 where φ > 1 is the symmetric elasticity of substitution across goods. The corresponding consumption-based price index, Pt , is given by: 1 1 1−φ 1−φ Pt = Pt (i) di (2) 0 Ricardian agents smooth their consumption, Cat , over time and thus maximize the life- ∞ 1−γc t (Cat ) time utility function E0 t=0 β , where γ is the risk aversion parameter and β is 1−γc the discount factor. Utility maximization is subject to the following budget constraint: 2 2 ψ At+1 ψ A∗t+1 At+1 + St A∗ t+1 + Pt + St Pt∗ + (1 + τtc )Pt Cat 2 Pt 2 Pt∗ ∗N ∗ f = (1 + iN A i t )At + (1 + it )At St + Pt (Tt + Tt + Tt ) Ricardian agents hold domestic assets At (denominated in domestic currency) on which they N ∗ receive the nominal interest rate it and foreign assets At+1 (denominated in foreign currency) ∗N on which they receive the interest rate it . Assets are subject to quadratic adjustment costs, measured by the parameter ψ . These costs are paid to nancial intermediaries whose only function is to collect these transaction fees and rebate the revenue to households in lump-sum c fashion in equilibrium. Ricardian households pay a consumption tax τt on their consumption Cta . St is the nominal exchange rate. Moreover, TtA is a lump-sum rebate of costs of adjusting i f asset holdings from the intermediaries to which it is paid and Tt and Tt are a lump-sum 78 rebate of prots from intermediate and nal goods production. At+1 A∗ If we denote Pt = a t+1 and t+1 Pt∗ = a∗ t+1 , we can re-write the budget constraint in real terms: ψ ψ ∗ 2 at+1 + Qt a∗ (at+1 )2 + Qt t+1 + a + (1 + τtc )Cat 2 2 t+1 (1 + iN t ) (1 + i∗N t ) f = at + ∗ Qt a∗ A i t + Tt + Tt + Tt 1 + πt 1 + πt where πt is the ination rate and 1 + πt = Pt Pt−1 . The term Qt = St Pt∗ /Pt stands for the real (1+iN t ) exchange rate. If we dene the domestic and foreign gross real interest rates as 1+ it = 1+πt (1+i∗N t ) and 1 + i∗ t = 1+πt ∗ , we can re-write the budget constraint as: ψ ψ 2 at+1 +Qt a∗ t+1 + (at+1 )2 + Qt a∗ c ∗ ∗ A i t+1 +(1+τt )Cat = (1+it )at +(1+it )at Qt +Tt +Tt +Tt f (3) 2 2 7 We assume that Ricardian households are rms' owners. 8 The denition of this set of lump-sum rebate of costs and prots is the same as in Cacciatore and Ghironi (2015) and hence we refer to their paper for a complete derivation of these variables. The only dierence in our model concerns the lump-sum rebate of prots from intermediate goods, which is dened wF t wIt as: Tti = Pt φt ZF t lF t − Pt lF t − Pt lIt − κF VF t − κI VIt . 5 where it and i∗ t are respectively the real interest rates on domestic and foreign assets. The Euler equations for domestic and foreign asset holding are respectively: −γc c Cat+1 1 + τt (1 + ψat+1 ) = (1 + it+1 )βEt −γc (4) Cat 1 + τtc+1 −γc c Cat+1 Qt+1 1 + τt (1 + ψa∗ t+1 ) = (1 + i∗ t+1 )βEt −γc (5) Cat Qt 1 + τtc+1 On the other hand, non-Ricardian households do not have access to nancial markets and hence they can nance their consumption needs either though labor income (wF t if they supply labor to the formal sector and wIt if they supply labor to the informal sector) or through unemployment benets (bt ) if they do not work. The following equations dene non-Ricardian agents' consumption depending on whether they work in the formal sector, or they work in the informal sector, or they are unemployed: (1 − τtw ) CF t = wF t lF t (6) (1 + τtc ) wIt CIt = lIt (7) (1 + τtc ) bt Cut = (1 − lt ) (8) (1 + τtc ) w The payroll tax on employees, τt , is borne only by non-Ricardian agents employed in the formal sector. Total labor supply, lt , is the sum of labor supplied by non-Ricardian households in the formal and informal sectors, i.e. lt = lF t + lIt . In equilibrium, aggregate unemployment is given by: Ut = 1 − lF t − lIt (9) Total consumption Ct is dened as the weighted sum of consumption of Ricardian house- holds (Cat ) and non-Ricardian households working in the formal sector (CF t ), in the informal sector (CIt ) or unemployed (Cut ): Ct = ωCat + (1 − ω )(CF t + CIt + Cut ) (10) where ω is the share of Ricardian households. 2.2 Production There are two vertically integrated production sectors. In the upstream, in both the formal and the informal sector, intermediate goods are produced in perfect competition using only labor. Intermediate goods are then sold to nal good producers. In the downstream, each sector i is populated by a representative monopolistically competitive multi-product rm, which uses intermediate goods as inputs to produce dierentiated varieties. In equilibrium, some of these varieties are exported while others are sold only on the domestic market. 6 2.2.1 Intermediate goods We assume a unit mass of intermediate good producers, which operate both in the formal and informal sectors. Both sectors are subject to search and matching frictions as in the Diamond-Mortensen-Pissarides framework. Unemployed agents search for a job in both sectors and search eorts are endogenous. Wages are set through an individual bargaining process. We assume a constant-return-to-scale matching technology in each sector j, for j = F, I , where F and I refer respectively to the formal and the informal sector. The matching technology converts aggregate unemployed workers, Ut , and aggregate vacancies, Vt , into aggregate matches, Mt . The matching rate in each j sector is: Mjt = χj (ejt Ut )1−ε Vjt ε (11) where Ut is the total number of unemployed workers and Vt is the number of vacancies. The parameters χ and measure respectively the matching eciency and the matching function elasticity, with χ>0 and 0< < 1. Let ejt denote search eorts for the job type j when agents are unemployed. The job lling rate, qt , is: 1− Mjt ejt Ut qjt = = χj (12) Vjt Vjt The job nding rate, ι is: Mjt Vjt ιjt = = χj ejt (13) Ut ejt Ut As in Krause and Lubik (2007), we assume that newly created matches become productive only in the next period. The law of motion of employment, ljt , is: ljt = (1 − λj )ljt−1 + qjt−1 vjt−1 (14) where λj ∈ (0, 1) is the exogenous separation rate and vjt is the number of vacancies posted by the rm in period t. In equilibrium vjt = Vjt . Firms, both in the formal and informal sector, hire labor lt to produce an intermediate good yjt according to the following technology : Int yjt = Zjt ljt ∀ j = F, I (15) where Zjt is an exogenous technology term which follows an autoregressive process AR(1): ∗ log Zjt = φZ 1 log Zjt−1 + φZ 2 log Zjt−1 + Z jt . (16) 7 In both sectors j = F, I , intermediate rms choose the number of vacancies, vjt , and employment, ljt , to maximize the discount value of their prots: ∞ uC,t f E0 βt ϕt Zjt ljt − wjt ljt (1 + τjt ) − κj vjt (17) t=0 uC,0 subject to the law of motion for labor: ljt = (1 − λj )ljt−1 + qjt−1 vjt−1 , where ϕt is the real price at which intermediate goods producers sell their goods to nal good producers and it 9 is expressed in units of consumption ; wF t is the wage paid to workers in the formal sector (lF t ), while wIt is the wage paid to workers in the informal sector (lIt ). In both sectors, intermediate good producers incur a cost of κj units of consumption per vacancy posted f vjt .The term τjt represents a payroll tax on employers. These taxes are paid only by rms f f operating in the formal sector. Hence τF t > 0, whereas τIt = 0. The rst order conditions (hereafter, FOCs) on vjt and ljt in the formal and informal sector are respectively: κj = Et [βt,t+1 µjt+1 ] (18) qjt f µjt = ϕjt Zjt − wjt (1 + τjt ) + Et [βt,t+1 (1 − λj )µjt+1 ] (19) where µjt is the Lagrangian multiplier for labor adjustment and measures the current value of an additional worker. Combining both FOCs leads to the job creation conditions in both sectors: κF κF f = Et βt,t+1 (1 − λF ) + ϕt+1 ZF t+1 − wF t+1 (1 + τF t+1 ) (20) qF t qF t+1 κI κI = Et βt,t+1 (1 − λI ) + ϕt+1 ZIt+1 − wIt+1 (21) qIt qIt+1 uC,t+1 where βt,t+1 ≡ β uC,t is the one period ahead stochastic discount factor. For both the formal and the informal sector, the job creation conditions state that, in equilibrium, the vacancy creation cost incurred by the rm per current match is equal to the expected discounted value of the vacancy creation cost per future match, further discounted by the probability of current match survival 1 − λ, plus the prots from the match at time t. Prots from the match take into account the future marginal revenue product from the match and its wage cost. Wages Nominal wages are set through an individual Nash bargaining process. In each t period and in both sectors J = F, I , the real value of an existing, productive match for a producer, Jt , is the sum of the marginal product of the match (ϕt Zjt ) and the expected discounted continuation value of the match (Et βt,t+1 (1 − λj )Jjt+1 ), net of the wage bill: f Jjt = ϕt Zjt − wjt (1 + τjt ) + Et βt,t+1 (1 − λj )Jjt+1 (22) 9 Firms are owned by households and uC,t is the marginal utility of consumption. This ensures that rst order conditions are measured in the same units. 8 The worker's value of being matched, in both the formal and informal sector, is given by the sum of real wage received and the expected discounted future value of being matched by the rm: w (1 − τjt ) Wjt = wjt + Et {βt,t+1 [(1 − λj )Wjt+1 + λj Uu,t+1 ]} (23) (1 + τtc ) The expected future value of being matched by the rm (the last term on the right-hand side of Eq.(23)) is a weighted average of probability 1−λ that the match will survive or the probability λ that the worker will become unemployed. The value of being unemployed is dened as: bt e1+ Ft e1+ Ut = −ϑ − ϑ It + Et {βt,t+1 [ιF t WF t+1 + ιIt WIt+1 +(1 − ιF t − ιIt )Uu,t+1 ]} (24) (1 + τtc ) 1+ 1+ e1+ jt where ϑ 1+ is a convex search cost and is the elasticity of disutility of searching. Therefore, 10 the value of being unemployed is the sum of unemployment benets  net of search costs  and the expected discounted future value of future states, where ιF t and ιIt are the probability of becoming employed respectively in the formal or informal sector. We dene worker's surplus Hjt ≡ Wjt − Ut . The worker surplus in the formal and informal sector is given by: w (1 − τjt ) bt e1+ Ft e1+ HF t = wjt − −ϑ − ϑ It + (1 − λF − ιF t − ιIt )Et (βt,t+1 HF t+1 ) (1 + τtc ) (1 + τtc ) 1+ 1+ (25) wIt bt e1+ Ft e1+ It HIt = − −ϑ −ϑ + (1 − λI − ιF t − ιIt )Et (βt,t+1 HIt+1 ) (26) (1 + τtc ) (1 + τtc ) 1+ 1+ η 1−η Nash bargaining maximizes the joint surplus Jjt Hjt with respect to wjt , where Hjt and Jjt stand for surpluses respectively for workers and rms and the parameter η measures the bargaining power of rms. The FOC implies: ∂Jjt ∂Hjt ηHjt + (1 − η )Jjt =0 (27) ∂wjt ∂wjt w 1−τjt ∂Jjt f ∂Hjt where ∂wjt = −(1 + τjt ) and ∂wjt = 1+τtc . Hence, the sharing rule can be rewritten in the following form: w f 1 − τjt (1 + τjt )ηHjt = (1 − η )Jjt (28) 1 + τtc 10 We assume that the informal sector does not allow the worker to be eligible for the unemployment benets. Given that we have a representative unemployed worker, we set an average unemployment benets, bt = lF t /(lF t + lIt )bWF t , where the parameter b is the replacement rate and measures benet generosity by comparing unemployment benets received when not working to wages earned when employed. 9 The bargained wage satises the following condition, respectively in the formal and informal sector: η bt e1+ Ft e1+ wF t = w c −ϑ − ϑ It 1 − τF t (1 + τ t ) 1 + 1+ f w 1−η 1 + τF t 1 − τF t+1 + f ϕt ZF t + Et βt,t+1 JF t+1 (1 − λF ) − (1 − λF − ιF t ) f w 1 + τF t 1 + τF t+1 1 − τF t (29) bt e1+ Ft e1+ It wIt = η − ϑ − ϑ + (1 − η ) [ϕt ZIt + ιIt Et (βt,t+1 JIt+1 )] (30) (1 + τtc ) 1+ 1+ Wages are a linear combination  determined by the bargaining power parameter η  of worker's outside option and the marginal revenue product generated by the worker plus the expected discounted continuation value of the match to the rm. For high values of η, the bargaining power of rms is higher and the portion of the net marginal revenue product and continuation value to the rm appropriated by workers as wage payments is smaller, hence the outside option becomes more relevant. Optimal search intensities are given by ∂Ut /∂ejt = 0, which yields: ∂ιjt ϑejt = Et (βt,t+1 Hjt+1 ) (31) ∂ejt 1−η VF t 1 − τtw +1 κF ϑeF t = χF (32) η eF t Ut (1 + τtf+1 )(1 + τtc+1 ) qF t 1−η VIt 1 κI ϑeIt = χI (33) η eIt Ut 1 + τtc+1 qIt This set of equations shows that search eorts are increasing in market tightness (Vjt /Ujt ) 1−τt w +1 and decreasing in taxes. We dene the tax wedge as T WF t = f c ) in the formal (1+τt+1 )(1+τt +1 1 sector and T WIt = c 1+τt in the informal sector. Equations above show that the higher the +1 tax wedge, the lower the search eort. However, the tax wedge is not symmetrical between sectors and hence the incentive to search for an informal job are reduced only by an increase c f w in the consumption tax, τt , but they are not aected by changes in payroll taxes, τt and τt . 2.2.2 Final goods In this subsection variables denoted by the letter d refer to a country's own goods consumed or produced domestically, whereas x refers to quantities and prices of exports. Producer i is a multi-product rm that produces a set of dierentiated product varieties, indexed by ω , y (ω, i), which is dened over a continuum Ω: θ ∞ θ −1 θ −1 Yt (i) = yt (ω, i) θ dω (34) ω ∈Ω 10 where θ >1 is the symmetric elasticity of substitution across varieties. To save notation, from now on, we omit the index i, since consumption-producing sectors are symmetric in the economy. We dene Pty , the cost of the product bundle Yt : 1 ∞ 1−θ Pty = py t (ω ) 1−θ dω (35) ω ∈Ω where py t (ω ) is the nominal marginal cost of producing variety ω. To create a new variety ω, each retailer needs to create a new plant, facing a sunk investment, fe,t , denominated in units of intermediate input. Each plant produces using dierent technologies indexed by relative productivity z (ω ), which is drawn from a common distribution G(z ) with support on [zmin , ∞). For the sake of simplicity, from now on we omit ω. This relative productivity level remains xed thereafter. Productivity level of foreign plants are drawn from an identical distribution. Each plant uses intermediate inputs to produce its dierentiated product variety, facing the real marginal cost: py t (z ) ϕt ϕz,t ≡ = (36) PT z The number of products created and commercialized by each retailer is endogenous. At each point in time, only a subset of varieties Ωt ⊂ Ω is actually available to consumers. Therefore, at time t, each Home retailer commercializes Nd,t varieties and creates Ne,t new products that will be available for sale at time t + 1. New and incumbent plants can be hit by a "death" shock with probability δ ∈ (0, 1) at the end of each period. The law of motion for the stock of producing plants is: Nd,t+1 = (1 − δ )(Nd,t + Ne,t ) (37) where δ is the rm's exit rate. When serving the foreign market, each retailer faces per-unit iceberg trade costs, τt > 1, as well as xed export costs, fx,t paid for each exported product and denominated in units of intermediate input. We dene total xed costs ¯ f x,t = fx,t Nx,t , where Nx,t denotes the number of product varieties exported abroad. If xed export costs ¯ are absent (f x,t = 0), each producer would nd it optimal to sell all its product varieties both domestically and abroad. Fixed export costs imply that only varieties produced by plants with suciently high productivity (above a cuto level zx,t , determined below) are exportable. We dene two special average productivity levels (weighted by relative output shares): an average zd for all producing plants and an average zx,t for all exporting plants: 1 1 θ −1 ∞ ∞ θ −1 1 zd = z θ−1 dG(z ) zx,t = z θ−1 dG(z ) zmin 1 − G(zx,t ) zx,t 11 1 We assume that G(·) is Pareto with shape parameter kp > θ − 1.11 As a result, zd = κ θ−1 zmin 1 and zx,t = κ θ−1 zx,t , where κ = kp /[kp − (θ − 1)]. The share of exporting plants is given by: −kp zmin kp Nx,t = [1 − G(zx,t )]Nd,t = κ θ−1 Nd,t (38) zx,t The real costs of producing the bundles Yd,t and Yx,t are respectively: y y Pd,t 1 1−θ ϕt Px,t 1 1−θ ϕt = Nd,t , = Nx,t (39) Pt zd Pt zx,t The nal producer determines Nd,t+1 and the productivity cuto zx,t to minimize the present discount value of costs: ∞ y y Pd,s Px,s Ns+1 βt,s Yd,s + τs Yx,s + − Ns fe,s ϕs + Nx,s fx,s ϕs (40) s=t Ps Ps 1−δ 1 subject to (38), (39), and zx,t = κ θ−1 zx,t . The FOC with respect to zx,t yields: y Px,t Yx,t (θ − 1)kp τt = fx,t ϕt (41) Pt Nx,t kp − (θ − 1) In equilibrium, the marginal revenue from adding a variety with productivity zx,t to the export bundle has to be equal to the xed cost. Thus, varieties produced by plants with productivity below zx,t are distributed only in the domestic market. The composition of the traded bundle is endogenous and the set of exported products uctuates over time with changes in the protability of export. The FOC with respect to Nd,t+1 determines product creation:   ϕt+1 fe,t+1 − N x,t+1 f Nd,t+1 x,t+1 ϕt fe,t = (1 − δ )βt,t+1  y Pd,t+1 Yd,t+1 y Px,t  (42) 1 +1 Yx,t+1 Nx,t+1 + θ− 1 Pt+1 Nd,t+1 + τt+1 Pt+1 Nx,t+1 Nd,t+1 In equilibrium, the cost of producing an additional variety, ϕt fe,t , must be equal to its expected benet, which includes expected savings on future sunk investment costs augmented by the marginal revenue from commercializing the variety, net of xed export costs, if it is exported. −kp 11 Hence, G(x) = z . zmin 12 Domestic and export prices Let Pd,t and Px,t be the price of the product bundle Yd,t and Yx,t . Each nal producer faces the following domestic and foreign demand for its product bundles: −φ −φ Pd,t Px,t Yd,t = YtC , Yx,t = YtC ∗ (43) Pt Pt∗ where YtC and YtC ∗ stand for aggregate demands of the consumption basket in the domestic and foreign country. The elasticity of substitution across sectoral bundles for the aggregate demand, φ > 1, is equal to the elasticity of substitution for the consumption basket, al- though aggregate demand in each country includes sources other than consumption. This assumption ensures that the consumption price index for the the consumption aggregator is also the price index for aggregate demand of the basket. We assume producer currency pricing (PCP): nal producers set the price of the product h bundle, Pd,t , and the the price of the export bundle, Px,t , in their own domestic currency, letting the price in the foreign market move with the nominal exchange rate, that is: Px,t = h τ Px,t /St . Because of xed export costs, the composition of domestic and export bundles is dierent, and hence producers face dierent marginal costs of producing these bundles. Therefore nal producers set two dierent prices for the Home and Foreign markets. The optimal price for domestic sales and exported sales satises respectively: y h y Pd,t φ Pd,t Px,t τt φ Px,t = , = (44) Pt φ − 1 Pt Pt Qt φ − 1 P t where Qt = St Pt∗ /Pt is the real exchange rate. 1 θ −1 We dene the average price of a domestic variety, ˜d,t ≡ Nd,t ρ (Pd,t /Pt ) and the average 1 price of an exported variety, ˜x,t ≡ Nx,t (Px,t /Pt∗ ). ρ θ −1 Combining the equations (39) and (44), we obtain the average price of a domestic and an exported variety, respectively dened as: φ ϕt φ τt ϕt ˜d,t = ρ , ˜x,t = ρ (45) φ−1z ˜d φ − 1 Qt z ˜x,t Finally, the average output of, respectively, a domestic and exported variety are dened as: θ −φ θ −φ y ˜− ˜d,t = ρ φ 1−θ C d,t Nd,t Yt , y ˜− ˜x,t = ρ φ 1−θ x,t Nx,t Yt C∗ (46) 2.3 Government In each period, we assume that government spending and unemployment benets are funded by taxation on consumption and wage income: Gt = τtc [ωCat + (1 − ω )(CF t + CIt + Cut )] + (τtw + τtf )wF t lF t − bt Ut (47) 13 2.4 Closing conditions Aggregate demand is the sum of private and public consumption and is dened as: YtC = ωCat + (1 − ω )(CF t + CIt + Cut ) + κF VF t + κI VIt + Gt (48) We assume that the cost of opening new vacancies are socially shared. Assets are in zero net supply, which implies the equilibrium condition: at+1 + a∗ t+1 = 0 (49) Net foreign assets are determined by: (at+1 − at ) + Qt (a∗ ∗ ∗ ∗ t+1 − at ) = it at + Qt it at + Qt Nx,t ρ ∗ ∗ ˜x,t − Nx,t ˜x,t y ρ ˜x,t y ∗ ˜x,t (50) where the last term in brackets represents the trade balance: T Bt = Qt Nx,t ρ ˜x,t − ˜x,t y ∗ ∗ ∗ Nx,t ˜x,t y ρ ˜x,t . 3 Calibration We calibrate the model using quarterly data from the U.S. and Brazilian economies. We believe that Brazil is an illustrative example of an emerging country which, starting form high level of informality in the late 1990s, has adopted a set of policy initiatives to facilitate 12 the move to formality. A rst program, called SIMPLES, was launched in 1996 and was followed by a second one, the SUPERSIMPLES program, in 2006. Since, in Brazil there is a strong correlation between size of company and prevalence of informality, these programs aimed at reducing the costs of formalization through a simplication and a reduction of tax 13 rates and tax regulations for Brazilian micro rms with no more than ve paid employees. Since the SUPERSIMPLES came into force in July 2007, some 9 million businesses have joined this system of taxation and the formal rate has increased by 11 percentage points (see 14 Fajnzylber et al. (2011)). In this section we discuss the calibration strategy. Broadly speaking, we choose some parameter values from the literature, while other parameters are set so to match macroeco- nomic series observed for the United States and Brazil. We assume that the two countries are asymmetric, hence some parameters describing labor and goods markets may dier across countries. Table 1 summarizes the asymmetric calibration. 12 See ILO (2014) for a discussion and an evaluation of other programs launched in emerging countries to move to formalization. 13 The SIMPLES program combined six dierent federal taxes and social contributions into a single monthly-based rate. The two reforms also reduced the tax burden considerably. 14 While Fajnzylber et al. (2011) nd very large eects of the SIMPLES program on formality rates, Monteiro and Assunção (2012) nd positive and signicant eects on formalization rates only among rms in the retailer sector. For a reconciliation of these two studies, see Piza (2016). 14 We set the discount factor β at 0.99, implying that the annual real interest rate is 4 percent. The value of the risk aversion parameter, γc , is equal to 2. Following Bernard et al. (2003), we set the elasticity of substitution across product varieties, θ, equal to 3.8. Following Ghironi and Melitz (2005), we set the elasticity of substitution across Home and Foreign goods, φ, equal to θ, and the dispersion of rm productivity kp equal to 3.4. We normalize zmin to 1. We set iceberg trade costs τ equal to 1.7, following the estimates of trade costs reported by Anderson and van Wincoop (2004). We calibrate the xed export costs fx so that the shares of exporting plants in the developed and emerging country are respectively equal to 21 percent and 18 percent, consistently with data reported in Bernard 15 et al. (2003) for the United States and in the World Bank Enterprise Survey for Brazil. To ensure steady-state determinacy stationarity of net foreign assets, we set the parameter ψ measuring asset adjustment costs equal to 0.0025 as in Ghironi and Melitz (2005). Following Ebell and Haefke (2009), we set entry costs, fe , so that regulation costs amount to 5.2 months of per capita output. To pin down the rm exit rate δ , we target the portion of worker separation due to rm exit equal to 30 percent in the United States and to 37 percent in Brazil: these values fall within the range of estimates reported by Haltiwanger et al. (2008). Empirical evidence indicates that informal rms are less productive than formal ones. We normalize the productivity parameter in the informal sector to unity and we assume that 16 the productivity in the formal sector is 30 percent higher than in the informal sector. Regarding the parameters specic to the search and matching framework, the gross re- placement rate for unemployment benets b in the formal sector is set to 13 percent for the United States and 15.2 percent for Brazil. The parameter measuring rms' bargaining power, η, is equal to 0.4, as estimated by Flinn (2006). The elasticity of the matching func- tion ε is equal to 0.4, so that it falls within the range of estimates reported by Petrongolo and Pissarides (2006) and the Hosios condition holds. We set the costs of vacancy posting (κF and κI ), matching eciency (χF and χI ) and exogenous separation rate (λF and λI ) in the formal and informal sectors so to match the underlying structure of the two countries, with the values of steady-state ratios summarized in Table 1. We choose a calibration based on the long-run averages (1992-2017) from ILO data. Steady-state unemployment rates are respectively 6 percent and 8.7 percent in the United States and Brazil, while the ratio of informal employment to total employment is respectively 7 percent and 30 percent in the 17 United States and Brazil. This calibration yields an informal wage gap (i.e. dierence between wages for formal and informal workers) equal to 66 percent in the United States and 11 percent in Brazil. This latter value is very close to estimates in Bargain and Kwenda (2010, 2014) who conclude that earning dierentials driven by the informal wage penalties are quite modest in Brazil and remain below 10 percent all along the distribution. Labor market regulations and high employer costs attached to formal employment in Brazil may simultaneously explain the large extent of informal work and the relatively modest informal 15 As a caveat, we point out that the World Bank Enterprise Survey covers only rms of the formal private sector with ve or more employees. Hence informal and micro rms are excluded from the sample. 16 This assumption allows us to reproduce a wage premium equal to 30 percent as in Charlot et al. (2015). 17 We use vulnerable employment as a proxy for informal employment. 15 wage gap. Firms tend to recoup high employers' payroll taxes paid to hire formal workers, which could partly explain low informal wage gaps. In Brazil informal wage penalties may only partly be related to the rm size eect, since many informal workers are to be found in large formal rms. Finally, we set the initial value of tax rates at their respective steady-state levels. The United States employs a retail sales tax rather than a value added tax (VAT) as the principal consumption tax. The retail sales tax in the United States is not a federal, but it is a tax imposed at the state and local government levels. The total tax rate ranges between 0 percent (e.g. in Delaware, Oregon, New Hampshire, Montana) and 13.5 percent (in Alabama). We c decide to set τ for the United States at the average rate, 7.8 percent. Brazil operates a multiple rate system with ICMS (Imposto de Circulação de Mercadorias e Serviços) tax levied at the state level. The standard state rate of ICMS is 17 percent (18 percent in São Paulo, Minas Gerais and Paraná and 19 percent in Rio de Janeiro). Therefore, for Brazil we c set τ equal to 17 percent. The personal income tax rate ranges between 0 percent and 37 percent in the United States and between 0 percent and 27.5 percent in Brazil. We choose w the average value of the personal income tax rate and we set τ equal to 18 percent for the United States and 14 percent for Brazil. In the United States, the social security tax rate is 12.4 percent (6.2 percent on employees and 6.2 percent on employers). On top, there is a tax of 2.9 percent (half imposed on employer and half withheld from the employee's pay) of all wages for Medicare. In Brazil, the employer's contribution is determined at the rate of approximately 20 percent of salary to be paid to the National Institute of Social Security (Instituto Nacional do Seguro Nacional, INSS). On top, the FGTS is the Fundo de Garantia por Tempo de Serviço which is the Employee Indemnity Guarantee Fund and an employee compulsory fund. All Companies are obligated to deposit the FGTS contribution into their employers account. The tax corresponds to an 8 percent rate on top of the gross salary. Since in our model we consider only the share of payroll taxes paid by employers, we set the f steady-state payroll tax rate, τ equal to 7.65 percent for the US and 28 percent for Brazil. 16 Table 1: Calibration Targets and parameters Notation Developed Emerging Source Calibration targets Formal employment lF /(lF + lI ) 93% 70% ILO, Trends Econometric Models Informal employment lI /(lF + lI ) 7% 30% ILO, Trends Econometric Models Unemployment rate U 6% 8.7% ILO, Trends Econometric Models Share of exporting rms Nx /Nd 21% 18% World Bank and Bernard et al. (2003) Final good Market Sunk entry costs fe 0.4 0.4 Ebell and Haefke (2009) Fixed export costs fx 0.0062 0.0090 Calibration targets Iceberg trade costs τ 1.7 1.7 Anderson and van Wincoop (2004) Pareto shape κp 3.4 3.4 Ghironi and Melitz (2005) Plant exit δ 0.026 0.026 Haltiwanger et al. (2008) Elasticity of substitution θ=φ 3.8 3.8 Bernard et al. (2003) Taxation Consumption tax τc 7.8% 17% Income tax τw 18% 14% Payroll tax τf 7.65% 28% Labor market Bargaining power η 0.4 0.4 Flinn (2006) Matching function elasticity ε 0.4 0.4 Petrongolo and Pissarides (2006) Vacancy costs, formal κF 2.5 2.5 Calibration targets Vacancy costs, informal κI 1.5 1.5 Calibration targets Matching eciency, formal χF 0.30 0.28 Calibration targets Matching eciency, informal χI 0.35 0.38 Calibration targets Separation rate, formal λF 0.032 0.055 Calibration targets Separation rate, informal λI 0.27 0.15 Disutility of search, scale ϑ 2 2 Disutility of search, elasticity 1.3 1.3 Unemployment benets, formal b 13 15.2 Aleksynska and Schindler (2011) Other parameters Risk aversion γc 2 2 Discount factor β 0.99 0.99 Bond adjustment cost ψ 0.0025 0.0025 Ghironi and Melitz (2005) 17 4 The impact of trade liberalization 18 In this section we introduce a trade shock which simulates the eects of trade liberalization. Since trade liberalization has often been blamed to be biased toward informality and to favor skilled/high-income workers, we believe that a trade shock is suitable to illustrate the possible 19 role of taxation in supporting labor formalization and achieving higher equality. The conventional view posits that trade liberalization causes an expansion of labor in- formality. However, the mechanism through which trade aects workers in the presence of informality is not clear. Intuitively, on the one side trade liberalization induces lower- productivity formal rms to switch to the informal sector to remain protable. The inci- dence of informality will increase accordingly. On the other side, trade liberalization induces lower-productivity informal rms to exit the market and hence the incidence of informality will decrease. The net eect on informality remains ambiguous. According to Bacchetta et al. (2009), trade expansion has not led to a corresponding improvement in working condi- tions and living standards for many. In many developing economies job creation has mainly taken place in the informal economy. The empirical literature provides mixed evidence on the eects of trade liberalization on informality, most likely because these eects are country- and/or industry-specic. Therefore, the relationship between trade liberalization and infor- mality strictly depends on the data underlying the empirical analysis. Some papers nd little or no eect of trade liberalization on informality (e.g. Goldberg and Pavcnik (2003), Menezes-Filho and Muendler (2011), Bosch et al. (2012)), whereas some others nd signif- icant eects of trade liberalization on informality. According to some studies trade reduces informal employment. Among these, Aleman-Castilla (2006) nds that a decline in US tar- is reduces informality in Mexico especially in export-oriented sectors. Currie and Harrison (1997) reach similar conclusions for government-owned rms in Morocco. Conversely, some other studies reach opposite conclusions and state that trade liberalization is associated with an increase in the share of informal workers (e.g. Ponczek and Ulyssea (2015) for Brazil and Acosta and Montes-Rojas (2014) for manufacturing rms in Argentina). Finally, some studies provide mixed evidence. Among these, Fugazza and Fiess (2010) show that macro-founded data tend to support the conventional view according to which trade liberal- ization causes a rise in informality, while micro-founded data do not. Using Brazilian data, Paz (2014) nds that while a cut in trading partner import taris decreases the share of domestic informal employment, a cut in domestic import taris has the opposite eect. Concerning the distributional eects of trade liberalization, the empirical literature pro- vides mixed evidence. On the one side, trade liberalization is deemed to have boosted the demand of skilled workers and hence triggered an increase in the relative wage of skilled 18 We acknowledge that the calibrated parameters already reect the eects of past liberalization in Brazil. By introducing a trade shock, we simulate the eects of possible further opening of the economy. 19 We consider a deterministic (perfect foresight), permanent reduction of policy parameters. Given the large size of the shocks, transition dynamics from the initial equilibrium to the nal equilibrium are found by solving the model as a nonlinear forward looking deterministic system using a Newton method. This method solves simultaneously all equations for each period. 18 to unskilled workers, the so called skill premium. As a consequence, income inequality has widened (see Epifani and Gancia (2006), Matsuyama (2007), Verhoogen (2008), Goldberg and Pavcinik (2007) and the literature mentioned herein). On the other side, other studies nd that trade liberalization reduces the skill premium and hence inequality especially in middle and low-income countries (see McCaig (2011) for Vietnam, Zhang and Wan (2006) for China, Amiti and Cameron (2012) for Indonesia, Robertson (2005) for Mexico, Gonzaga et al. (2006) for Brazil, Kumar and Mishra (2008) for India). In our set-up, trade liberalization is captured by a reduction in xed export costs in both countries and it is modeled in the following way: in a rst phase, the "Home" country, which is the developed country (i.e. the United States), cuts its per-unit iceberg trade costs (i.e. τt decreases from 1.7 to 1.5). This process starts at the beginning of the simulation period and ends 70 quarters later (i.e. 17.5 years later). The cut in iceberg trade costs gives a competitive advantage to the developed country. In a second phase, which starts 5 years later (i.e. 20 quarters later), the emerging country (i.e. Brazil) experiments the same decline in its own iceberg costs. Hence, 22.5 years after the initial reduction of trade costs observed in the developed country, iceberg costs in the emerging country will have converged to those observed in the developed country. At this third phase, the two countries benet from the same reduction in trade costs and trade liberalization becomes symmetrical. For the sake of clarity, we rst analyze the dynamics in the developed country and then in the emerging country. We discuss the dynamics both in the short term, i.e. before than the emerging country benets from trade liberalization, and in the medium to long term, i.e. when both countries can take advantage for the trade cost reductions. Simulations for the good sector and prices are displayed in Figure 1, while for the labor market are displayed in Figure 2 and Figure 3. Responses to the trade shock are represented by the blue solid lines. The impact on the goods sector. In the rst phase of trade liberalization, the devel- oped country cuts its trade costs, but the emerging country still does not benet from new technologies allowing it to reduce its trade costs. In this phase, lower trade costs in the developed country allow exporters to have higher prots. Trade translates into increased protable opportunities for exporting rms, which induces more rms to enter the export market. These rms face lower costs and hence increase their labor demand, which ulti- mately leads to higher real wage. Higher production costs, in turn, brings down the prots of the least productive rms and hence, at a second stage, rm entry is reduced. Notice that lower-productivity rms do not export and produce only for the domestic market. Hence, as it is shown in the Figure 1, the number of rms producing only for the domestic market in the developed country declines, but at the same time the number of exporters in this country increase (the export-cuto decreases). A higher proportion of exporting rms in the developed country leads to higher average quality of goods and higher productivity, as indicated also in other studies (e.g. Aleman- Castilla (2006)). In the short run, the emerging economy does not observe a decline of trade costs. In- 19 Figure 1: The nal good sector Domestic producers-H Exporters-H Productivity-H 0 40 0.4 -1 30 0.3 -2 20 0.2 -3 10 0.1 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Price in H -H Price in F -H Real exch. rate -0.2 1.2 -0.1 -0.4 1 -0.6 -0.2 -0.8 0.8 0.6 -0.3 -1 -1.2 0.4 -0.4 -1.4 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Domestic producers-F Exporters-F Productivity-F 40 -1 0.6 30 -2 20 0.4 -3 -4 10 0.2 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Price in H -F Price in F -F 1.5 0.4 TL 1 0.2 0.5 TL&TR 0 20 40 60 80 100 120 20 40 60 80 100 120 Note: H and F indicate respectively the Home country (i.e. the developed country) which is represented by a solid line and the Foreign country (i.e. the emerging country) which is represented by a dotted line. The blue lines display the dynamics with only trade liberalization, and the red lines display the dynamics when the tax reform is implemented. 20 stead, higher home prices in the developed country, combined with the decrease in the real exchange rate, lead consumers in the developed country to redirect their demand toward their trade partner (i.e. the emerging economy). This increase in demand addressed to emerging economy motivates more exporting rms in the emerging country. This, in turn, leads to a rise in input demand, and thus to a rise in the production costs (see Figure 1). As a consequence, input demand and production costs increase, which ultimately reduce prots for low-productive domestic rms. As a consequence, the number of new rm entries: the number of rms (Nf = "Domestic producers - F") declines in the emerging country. In the medium run, trade liberalization also aects the emerging country, where iceberg costs also decline, although with a delay. Hence, higher rms' prots worldwide boost income and labor demand leading to higher wages. The increase in labor costs leads both economies to be more selective: the number of rms declines, but the share of exporting rms, which are more productive, increases. In the emerging economy, in the medium run, trade liberalization ultimately induces more rms to export, thereby increasing labor demand and real wages. As in the developed economy, this leads to high share of exporters and informality in emerging economy. In the long run, when the developed country has reached its long-run level of iceberg costs, in the emerging country trade expansion is still ongoing. In the emerging country, revenue growth is now driven by iceberg cost reduction which takes place only in the emerging country and still generates growth gains. Growth gains, in this phase, are obviously more modest than during the rst phase of trade expansion. The impact on prices and exchange rate. In the developed country, at each period, domestic market prices (ρd,t in the model notation and "Price in H - H" in the gures) and export prices (ρx,t in the model notation and "Price in F - H" in the gures) are given by the following equations: φ ϕt ρd,t = φ − 1 zd τt φ ϕt ρx,t = Qt φ − 1 zx,t Hence, the increase in input prices, ϕt , generated by the expansion in nal good producers' demand explains the rise in domestic market prices ρd,t . On the other hand, export prices ρx,t drops as trade liberalization, through the decline in trade costs τt , compensates the increase in input prices as well as the decline in productivity (zx,t ) of export rms. Finally, lower iceberg costs in the developed country leads to a decline in the real exchange rate (Qt ) underlining the gains in competitiveness of this country. ∗ In the emerging country, at each period, the price of domestic goods (ρd,t and "Price in H ∗ - F" in the gures) and the price of exported goods (ρx,t and "Price in F - F" in the gures) 21 are modeled as in the developed economy, in a symmetric way: φ ϕ∗ ρ∗ d,t = t φ−1z d ∗ φ ϕ∗ ρ∗ x,t = τt Qt t φ − 1 z x,t ∗ ∗ It is clear that the rise of input price causes the rise in the domestic price (ρd ). The increase ∗ ∗ of the input price (ϕ ) as well as the decline of productivity of exporters (zx ) raises the export price, even though the real exchange rate (Qt ) declines. In the medium run, the larger participation of the emerging country to the world trade stabilizes export prices in the developed country: the real exchange rate is more stable and the bias cost in favor of the developed country slows down (see Figure 1). In the long run, when the developed country has reached its long-run level of iceberg costs, in the emerging country trade expansion is still ongoing. Therefore, the emerging country still benets from decreasing iceberg costs. Hence, its competitiveness is restored and the real exchange rate increases (see Figure 1). The impact on the labor market. In the developed country, higher input prices for nal producers translate into higher marginal revenues for the intermediate good producers, and ultimately into higher wages. Figure 2 shows that labor demand increases in both the formal and the informal sector, driven by the increase in the price of intermediate goods sold to nal producers. A part of this increase in the job surplus is redistributed to workers via wage increases. Figure 2 shows that wages increase in both the formal and the informal sector. Given that these wage increases are driven by the rise in the price of intermediate goods in both sectors, they are similar across the formal and the informal sector and thus wage inequalities remain stable (see Figure 4 in Appendix B). Although employment increases in both sectors, in the informal sector the increase is relatively larger, due to lower labor costs, which ensures that more job vacancies are opened in the informal sector. Indeed, expanded job creation in the informal sector encourages unemployed agents to search for a job more intensively in this sector, thus reinforcing the sector's advantage in the hiring process (see Figure 3). Therefore, at the beginning of the process trade liberalization induces higher informality in the developed country, along with a reduction in unemployment (see Figure 3). Tightness in labor market increases in the emerging country, although for reasons dier- ent from those observed in the developed country, and consequently employment and wage rise. As in the developed country, lower labor costs in the informal sector favor this sector during the expansion (see Figure 2). Moreover, unemployment declines, while the share of informality goes up (see Figure 3). Note that the rise in informality is of small amplitude in the emerging economy. This is due to the initial share of informal employment. As the emerging economy has a larger share of informality, it causes a more negative congestion eect: the job lling rate falls more rapidly with vacancy postings. Hence, this curbs job openings in the informal sector. 22 In the medium run, the increasing participation of both countries in the world trade, by increasing incomes and thus the demand for goods, boosts labor demand (see Figure 2) and reduces unemployment (see Figure 3). In the long run, when trade costs drop only in the emerging country, income growth generated by new exports is marginal: employment gains become smaller and smaller in both countries (i.e. developed and emerging) and both sectors (i.e. formal and informal). When iceberg costs converge to their long-term levels in both countries, variables converge towards the new steady-state levels. This phase is characterized by an over-adjustment, which is the result of vacancy-posting strategies adopted by rms (see Figure 2). As long as prot opportunities grow, there are strong incentives to post vacancies to benet from growth. This competition leads rms to over-hiring. Once growth falters, employment starts decreasing through the exogenous rate of destruction and the slowdown in new job opportunities. This process takes time and explains why, after the strong employment gains recorded during the period of trade expansion, both countries enter a phase characterized by a contraction on the labor market. Since the separation rate is higher in the informal sector than in the formal sector, the decline in employment is faster in the informal sector, which explains the rise in the share of formal employment in this phase of the long-term adjustment. To sum up, we observe that following a decline in trade costs wages increase in both the formal and the informal sector without changing the wage gap. Lower trade costs, although not harmful to equity, are biased toward labor informality. 5 Tax reform In order to reduce the increasing incidence of informality induced by trade liberalization, both countries should introduce incentives to develop businesses in the formal economy. An easy way to promote formal employment is to reduce the payroll tax paid by rms operating in the formal sector. Nevertheless, the cost of this policy is a reduction of public revenues which the government may use to nance public expenditures on social security. An alternative solution might be implementing a budget-neutral tax reform, consisting in increasing the consumption tax to fund the cut in payroll taxes. An advantage of this strategy is that the consumption tax has a larger base, it is easier to collect and more dicult to evade. This policy mix, called "social VAT", has been implemented in many European countries, for instance in Denmark in 1988, in Sweden in 1993, in Germany in 2006 and in France in 2012. In the rest of the paper, the tax reform is implemented in both countries at the beginning of their respective trade liberalization process. The tax reform is country-specic and budget- neutral. Given these constraints, the payroll tax is reduced from 8.0 percent to 5.8 percent with an increase in the consumption tax from 8 percent to 9.8 percent in the developed countries, whereas in the emerging country, the payroll tax is reduced from 28.0 percent to 24.0 percent with an increase in the consumption tax from 17.0 percent to 18.8 percent. Simulations for the goods sector are displayed in Figure 1, while for the labor market 23 Figure 2: The labor market Formal wage−H Formal employment−H Share of formal emp−H 0.04 3 0.02 TL 0 2 TL&TR 0.02 −0.02 1 0 −0.04 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Informal wage−H Informal employment−H Share of informal emp−H 1.4 0.6 0.6 1.2 0.4 0.4 1 0.2 0.2 0.8 0 0 0.6 0.4 −0.2 −0.2 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Formal wage−F Formal employment−F Share of formal emp−F 5 4 0.1 0.02 3 0 0.05 2 −0.02 1 0 −0.04 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Informal wage−F Informal employment−F Share of informal emp−F 1.5 0.2 0.05 1 0.1 0 0 0.5 −0.05 −0.1 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Note: H and F indicate respectively the Home country (i.e. the developed country) and the Foreign country (i.e. the emerging country). The blue lines display the dynamics with only trade liberalization, and the red lines display the dynamics when the tax reform is implemented. 24 Figure 3: Unemployment Unemployment−H Formal search effort−H Informal search effort−H TL 0.8 1 0 TL&TR 0.6 0.5 −0.5 0.4 0 0.2 −1 −0.5 0 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Unemployment−F Formal search effort−F Informal search effort−F 1 1 0 0.8 0.5 −0.5 0.6 0 −1 0.4 −1.5 0.2 −0.5 20 40 60 80 100 120 20 40 60 80 100 120 20 40 60 80 100 120 Note: H and F indicate respectively the Home country (i.e. the developed country) and the Foreign country (i.e. the emerging country). The blue lines display the dynamics with only trade liberalization, and the red lines display the dynamics when the tax reform is implemented. 25 they are displayed in Figure 2 and Figure 3. The scenario simulating the eects of the tax reform when the two countries are hit by a trade shock is represented by the red dotted lines. The impact on the nal good sector. Figure 1 depicts the eects of trade liberalization in the nal good sector when the government implements a budget-neutral tax reform. Taxation has no direct impact on the behavior of nal goods producers. The comparison with the pre-reform scenario (represented by the blue solid lines in Figure 1) points out that the dynamics of variables in the nal good sector remain unchanged because the main driver of both short-run and long-run changes in productivity and prices is trade liberalization. The tax reform only aects the distribution of jobs, across the formal and informal sector leaving 20 the aggregate demand of intermediate goods unchanged. This is due to the ambiguous eect of a budget-neutral tax reform on the tax wedge: on the one hand, it reduces the tax wedge by lowering the taxes paid by the employers, on the other hand it increases it by increasing the tax on consumption. The impact on labor markets. Figure 2 reports the eects of trade liberalization on labor markets when the government implements a budget-neutral tax reform. Recall that wages in both sectors are determined by the following equations: η b e1+ Ft e1+ 1−η ιF wF t = w c −ϑ − ϑ It + f ϕt ZF t + κF 1 − τF 0 (1 + τ1 ) 1+ 1+ 1+ τF 1 qF t b e1+ Ft e1+ It ιIt wIt = η c −ϑ −ϑ + (1 − η ) ϕt ZIt + κI (1 + τ1 ) 1+ 1+ qIt w where τF 0 is the tax paid by employees before the reform (indexed by 0). This tax rate remains unchanged, while the payroll tax paid by employers and the consumption tax jump f c instantaneously to their new post-reform values (respectively τF 1 and τ1 ). As observed for the baseline simulation without the tax reform (Figure 2, blue solid lines), wages increase in both sectors. However, when the tax reform is implemented, the increase in wages is more remarkable in the formal sector than in informal sector (Figure 2, red dotted lines). As a consequence, the wage gap between formal and informal workers is getting wider. Figure 4 shows that, before the tax reform, the wage gap between the formal and the informal sector was 65.7 percent in the advanced economy and 10.6 percent in the emerging country (see blue solid lines). After the reform, this gap rises to 69.2 percent in the advanced economy and to 14.4 percent in the emerging country (see red dotted lines). Widening wage gaps across the two sectors stem from the reduction of tax wedges, leading to a larger job surplus and thus higher wages. The tax reform also changes the sharing rule between rms and workers, at the advantage of the workers. The underlying mechanism is due to two channels: on the one hand, the drop in the tax paid by employers increases 20 To be more precise, changes in tax rates alter the equilibrium level of the production of intermediate goods. However, these changes have a second-order magnitude. 26 the share of productivity paid to employees in the formal sector. On the other hand, the increase in the consumption tax reduces the disposable wage. However, this moderation is proportional to the weight of the unemployment benets in the wage: as it is weak for workers in the formal sector, this wage moderation induced by the increase of the consumption tax is of small amplitude for the formal sector. The rst channel clearly dominates and leads to wage increases in the formal sector after the tax reform. Given that the search eort is endogenous, the tax reform also changes workers' reser- vation wage. Indeed, the cut in payroll taxes stimulates rms to open new vacancies in the formal sector, which in turn increases the chance for unemployed agents to nd a job in the formal sector. The optimistic job prospect in the formal sector encourages unemployed to focus their search eorts more on this sector. Hence, search eorts increase in the formal sector and decline in the informal sector (see Figure 3, red dotted lines). Overall, the tax 21 reform ultimately redirects the labor force toward formal employment. Figure 3 shows that, following the tax reform, unemployment increases on impact and in the short-run. The underlying reason is that benets from trade liberalization are gradual, while the tax reform is immediate: given the lack of attractiveness of the informal sector, search eorts  devoted to nd a job in the informal sector before the implementation of the tax reform  now decrease, leading to an increase of unemployment in the short run (see Figure 3, red dotted lines). At the beginning of the trade liberalization process, the marginal value of intermediate goods and workers' productivity, although higher, are not large enough to absorb the excess of unemployed workers who stop searching for an informal job. This explains why unemployment increases on impact and in the short-term especially in the emerging country, which is characterized by higher incidence of informality. To sum up, we observe that a tax reform switching the tax burden from payroll taxes to the VAT supports labor formality. Overall, output and economic eciency improve as the share of formality increases by 1 percentage point. However, as shown in Figure 4, these gains in formalization come at the cost of widening inequality between formal and informal workers. Our model estimates that the wage gap will rise from 65.7 to 69.2 percent in the advanced economy and from 10.6 to 14.4 percent in the emerging economy. Widening inequality may be attributed to the specic design of this tax reform, since the VAT has traditionally deemed to be regressive and hence harmful to equity. However, more recently some commentators and especially international organizations have pointed out that the VAT is not necessarily bad for redistribution. For instance, the distributional consequences of a tax reform switching the tax burden on the VAT have to be assessed in the perspective of the whole tax-spending system, of which the VAT is just one part (OECD (2010)). The VAT can still be progressive, if VAT revenues are used to nance benets targeted for poorer households. Notwithstanding these caveats, the empirical evidence (albeit limited) for a few 21 Similar conclusions are drawn in Antón (2014) who analyzes the eects of the 2012 tax reform in Colombia. He suggests that the reform would increase total employment by between 0.3 to 0.5 percent and formal employment by between 3.4 to 3.7 percent over the pre-reform scenario. In Brazil, tax cuts for small rms introduced by the reforms in 1996 and 2006 have led more than 9 millions of businesses into the formal sector. 27 developing countries nds that the VAT is not necessarily regressive. Empirical evidence from Bangladesh based on household income expenditure survey data nds that the VAT may have progressive elements (Faridy and Sarker (2011)). Therefore, as pointed out by the International Tax Dialogue (2013), the incidence of this regressivity is very country specic and generalizations can be misleading. Moreover, Ciminelli et al. (2019) nd that the labor market response does matter to assess the redistributive eects of the VAT. The VAT can reduce income inequality by triggering a positive labor supply channel. Higher indirect taxes increase the price of the consumption basket and create incentives for agents to increase their labor supply. This eect tends to be stronger for middle-aged women. 6 Conclusions In this paper, we show that trade liberalization boosts economic activity in both developed and emerging countries. However, we nd that trade liberalization is associated to higher informality, which ultimately implies less job security and lower employment quality. Policy makers should consider placing a high priority on promoting job quality and income equality. Policy interventions to curb informality should follow a comprehensive approach that rests on three pillars: increasing the benets of formality, decreasing the costs of formalization and improving enforcement methods. In this respect, we investigate whether taxation may smooth the way for formalization of jobs. To this purpose, we extend the Melitz (2003) model and develop a two-country DSGE model, featuring a developed and emerging (or developing) country as in Cacciatore and Ghironi (2015). In addition, we embed the tax system and the informal labor sector in the model with the aim to analyze whether taxation may correct the bias toward informality introduced by trade liberalization. We argue that a scal reform can mitigate these adverse eects of trade on the labor market. A Social VAT, switching the tax burden from payroll taxes paid by rms in the formal sector to the consumption tax, can increase the incentives to operate in the formal sector. Overall, output and economic eciency improve as the share of formality increases by 1 percentage point. However, these (small) gains in formalization come at the cost of widening inequality between formal and informal workers. Our model estimates that the wage gap will rise from 65.7 to 69.2 percent in the advanced economy and from 10.6 to 14.4 percent in the emerging economy. Our results suggest that (budget-neutral) tax reforms might play only a minor role in improving formalization rates in emerging countries. To reduce the incidence of informality, tax policy interventions should go hand in hand with other more eective spending and regulatory policies. Extending unemployment benets to all workers in the formal sector including those working part-time and/or on temporary contracts, could prevent the unem- ployed from looking for an informal job. Another step to enhance the quality of existing jobs is intensifying labor inspections in those sector where the incidence of informal work is higher. 28 References Acosta, P.; Montes-Rojas, G. 2014. Informal jobs and trade liberalisation in Argentina, in Journal of Development Studies, Vol. 50, No. 8, pp. 1104-1118. Adjemian, S.; Bastani, H.; Juillard, M.; Karamé, F.; Maih, J.; Mihoubi, F.; Perendia, G.; Pfeifer, J.; Ratto, M.; Villemot, S. 2011. Dynare: Reference Manual Version 4, Dynare Working Papers No. 1 (CEPREMAP). Aleksynska, M.; Schindler, M. 2011.  Labor market regulations in low-, middle- and high-income countries: A new panel database, IMF Working Paper No. 11/154 (In- ternational Monetary Fund). Aleman-Castilla, B. 2006. The eect of trade liberalization on informality and wages: Evidence from Mexico, CEP Discussion Paper No. 763 (London School of Economics). Amiti, M.; Cameron, L. 2012. Trade liberalization and the wage skill premium: Evidence from Indonesia, in Journal of International Economics, Vol. 87, No. 2, pp. 277-287. Anand, R.; Khera, P. 2016. Macroeconomic impact of product and labor market reforms on informality and unemployment in India, IMF Working Papers No. 16/47 (Inter- national Monetary Fund). Anderson, J. E.; van Wincoop, E. 2004. Trade costs, in Journal of Economic Literature, Vol. 42, No. 3, pp. 691-751. Antón, A. 2014. The eect of payroll taxes on employment and wages under high labor informality, in IZA Journal of Labor & Development, Vol. 31, pp. 1-23. Araujo, J.; Rodrigues, M. 2016. Taxation, credit constraints and the informal economy, in EconomiA, Vol. 17, pp. 43-55. Bacchetta, M.; Ernst, E.; Bustamante, J. 2009. Globalization and informal jobs in devel- oping countries. A joint study of the International Labour Oce and the Secretariat of the World Trade Organization, Geneva. Bargain, O.; Kwenda, P. 2014. The informal sector wage gap: New evidence using quantile estimations on panel data, in Economic Development and Cultural Change, Vol. 63, No. 1, pp. 117-153. Bargain, O.; Kwenda, P. 2010. Is informality bad? Evidence from Brazil, Mexico and South Africa, IZA Discussion Papers No. 4711. Batini, N.; Levine, P.; Lotti, E.; Bo, Y. 2011. Informality, frictions and monetary policy, University of Surrey School of Economics Discussion Papers No. 0711 . 29 Becker, D. 2018. Heterogeneous rms and informality: The eects of trade liberalization on labour markets, in Oxford Economic Papers, Vol. 70, No. 1, pp. 47-72. Bernard, A.; Eaton, J.; Jensen, B.; Kortum, S. 2003. Plants and productivity in interna- tional trade, in American Economic Review, Vol. 93; No. 4, pp. 1268-1290. Bosch, M.; Goñi-Pacchioni, E.; Maloney, W. 2012. Trade liberalization, labor reforms and formal-informal employment dynamics, in Labour Economics, Vol. 19, No. 5, pp. 653-667. Bosch, M.; Esteban-Pretel, J. 2015. The labor market eects of introducing unemployment benets in an economy with high informality, in European Economic Review, Vol. 75(C), pp. 1-17. Busato, F.; Chiarini, B. 2004. Market and underground activities in a two-sector dynamic equilibrium model, in Economic Theory, Vol. 23; No. 4, pp. 831-861. Cacciatore, M.; Ghironi, F. 2015. Trade, unemployment, and monetary policy, manuscript. Chacaltana, J.; Diaz, J. J.; Rigolini, J.; Ruiz, C. 2018. Pathways to formalization: Going beyond the formality dichotomy, World Bank Policy Research Working Paper No. 8551. Charlot, O.; Malherbet, F.; Terra, C. 2015. Informality in developing economies: Regula- tion and scal policies, in Journal of Economic Dynamics and Control, Vol. 51(C), pp. 1-27. Ciminelli, G.; Ernst, E.; Giuliodori, M.; Merola, R. 2019. The composition eects of tax- based consolidations on income inequality, in European Journal of Political Economy, Vol. 57(C), pp. 107-124. Coelho, D.; Corseuil, C. H.; Foguel, M. N. 2017. Employment growth of establishments in the Brazilian economy: Results by age and size groups, in Business Dynamics and Productivity (OECD Publishing, Paris). Colombo, E.; Menna, L.; Tirelli, P. 2018. Informality and the labor market eects of - nancial crises, DISEIS - Quaderni del Dipartimento di Economia internazionale, delle istituzioni e dello sviluppo dis1801, Università Cattolica del Sacro Cuore, Diparti- mento di Economia internazionale, delle istituzioni e dello sviluppo (DISEIS). Conesa, J. C.; Diaz-Moreno, C.; Galdon-Sanchez, J. E. 2002. Explaining cross-country dierences in participation rates and aggregate uctuations, in Journal of Economic Dynamics and Control, Vol. 26, pp. 333-345. Cook, D.; Nosaka, H. 2005. Dual labor markets and business cycles, Federal Reserve Bank of San Francisco Working Paper Series No. 2006-36. 30 Currie, J.; Harrison, A. 1997. Sharing the costs: The impact of trade reform on capital and labor in Morocco, in Journal of Labor Economics, Vol. 15, No. 3, pp. 44-71. Dellas, H.; Malliaropulos, D.; Papageorgiou, D.; Vourvachaki, E. 2017. Fiscal policy with an informal sector, Bank of Greece Working Papers No. 235. Ebell, M.; Haefke, C. 2009. Product market deregulation and the U.S. employment mira- cle, in eview of Economic Dynamics, Vol. 12, No. 3, pp. 479-504. Enste, D. H.; Schneider, F. 2000. Shadow economies: Size, causes, and consequences, in Journal of Economic Literature, Vol. 38, No. 1, pp. 77-114. Epifani, P.; Gancia, P. 2006. Increasing returns, imperfect competition, and factor prices, in The Review of Economics and Statistics, LXXXVIII, 4, pp. 583-598. Fajnzylber, P.; Maloney, W. F.; Montes-Rojas, G. W. 2011. Does formality improve micro-rm performance? Evidence from the Brazilian SIMPLES program, in Journal of Development Economics, Vol. 94, pp. 262-276. Faridy, N.; Sarker, T. K. 2011. Progressivity of Value Added Tax in developing countries: Empirical evidence from Bangladesh, in Asia-Pacic Tax Bulletin, Vol. 17, No. 3, pp. 185-191. Flinn, C. J. 2006. Minimum wage eects on labor market outcomes under search, matching, and endogenous contact rates, in Econometrica, Vol. 74, No. 4, pp. 1013-1062. Fugazza, M.; Fiess, N. 2010.  Trade liberalization and informality: New stylized facts, UN Policy Issues in International Trade and Commodities Study Series No. 43. Gasparini, L.; Tornarolli, L. 2007. Labor informality in Latin America and the Caribbean: Patterns and trends from household survey microdata, CEDLAS Working Paper. Ghironi, F.; Melitz, M. 2005. International trade and macroeconomic dynamics with heterogeneous rms, in The Quarterly Journal of Economics, Vol. 120, No. 3, pp. 865-915. Goldberg, P. K.; Pavcnik, N. 2003. The response of the informal sector to trade liberal- ization, in Journal of Development Economics, Vol. 72, No. 2, pp. 463-496. Goldberg, P. K.; Pavcnik, N. 2007. Distributional eects of trade liberalization in devel- oping countries, in Journal of Economic Literature, Vol. 45, No. 1, pp. 39-82. Gonzaga, G.; Filho, N. M.; Terra, C. 2006. Trade liberalization and the evolution of skill earnings dierentials in Brazil, in Journal of International Economics, Vol. 68, pp. 345-367. 31 Haltiwanger, J.; Scarpetta, S.; Schweiger, H. 2008. Assessing job ows across countries: The role of industry, rm size and regulations, NBER Working Papers No. 13920 (National Bureau of Economic Research, Inc.). Heintz,J.; Pollin, R. 2005. Informalization, economic growth, and the challenge of cre- ating viable labor standards in developing countries, in Rethinking Informalization: Poverty, Precarious Jobs and Social Protection, N. Kudva and L. Beneria (eds.) Cor- nell University Open Access Repository. Helpman, E.; Itskhoki, O.; Redding, S. 2010. Inequality and unemployment in a global economy, in Econometrica, Vol. 78, No. 4, pp. 1239-1283. Hsieh, C.-T.; Olken, B. A. 2014. The missing `missing middle' , in Journal of Economic Perspective, Vol. 28, No. 3, pp. 89-108. ILO 2013. Growth strategies and quality employment generation (International Labour Organization, Geneva). ILO 2014. Enterprise formalization: Fact or ction?  New ILO/GIZ study on what works to formalize enterprises (International Labour Organization, Geneva). International Tax Dialogue 2013. A decade of sharing tax experiences and knowledge: Key issues and debates in VAT, SME taxation and the tax treatment of the nancial sector. Jütting, J.; Parlevliet, J.; Xenogiani, T. 2008. Informal employment re-loaded, in IDS Bulletin, Vol. 39, No. 2, pp. 28-36. Krause, M. U.; Lubik, T. A 2007. The (ir)relevance of real wage rigidity in the New Keynesian model with search frictions, in Journal of Monetary Economics, Vol. 54, pp. 706-727. Kumar, U.; Mishra, P. 2008. Trade liberalization and wage inequality: Evidence from India, in Review of Development Economics, Vol. 12, No. 2, pp. 291-311. Matsuyama, K. 2007. Beyond icebergs: Towards a theory of biased globalization, in Review of Economic Studies, Vol. 74, No. 1, pp. 237-253. McCaig, B. 2011. Exporting out of poverty: Provincial poverty in Vietnam and U.S. market access, in Journal of International Economics, Vol. 85, No. 1, pp. 102-113. Melitz, M. 2003. The impact of trade on intra-industry reallocations and aggregate indus- try productivity, in Econometrica, Vol. 71, No. 6, pp. 1695-1725. Menezes-Filho, N. A.; Muendler, M. A. 2011. Labor reallocation in response to trade reform, NBER Working Paper No. 17372 (National Bureau of Economic Research). 32 Monteiro, J. C. M.; Assunção, J. J. 2012. Coming out of the shadows? Estimating the im- pact of bureaucracy simplication tax cut on formality in Brazilian microenterprises, in Journal of Development Economics, Vol. 99, pp. 105-115. OECD 2010. Tax policy reform and economic growth, Tax Policy Study No. 2 (OECD Publishing, Paris). OECD 2011. Divided we stand: Why inequality keeps rising (OECD Publishing, Paris). OECD 2015. Enhancing job quality in emerging economies, in OECD Employment Out- look 2015 (OECD Publishing, Paris). Orsi, R.; Raggi, D.; Turino, F. 2014. Size, trend, and policy implications of the under- ground economy, in Review of Economic Dynamics, Vol. 17, No. 3, pp. 417-436. Pappa, E.; Sajedi, R.; Vella, E. 2015. Fiscal consolidation with tax evasion and corruption, in Journal of International Economics, Vol. 96(S1), pp. 56-75. Paz, L. S. 2014. The impacts of trade liberalization on informal labor markets: A the- oretical and empirical evaluation of the Brazilian case, in Journal of International Economics, Vol. 92, No. 2, pp. 330-348. Petrongolo, B.; Pissarides, C. 2006. Scale eects in markets with search, in Economic Journal, Vol. 116(508), pp. 21-44. Piza, C. 2016. Revisiting the impact of the Brazilian SIMPLES program of rms' formal- ization rates, World Bank Policy Research Working Paper No. 7605 (World Bank). Poirier, A.; Trupkin, D. R. 2018. A model with labour market frictions and informality estimated for Argentina, manuscript. Ponczek, V.; Ulyssea, G. 2015.  Trade opening, enforcement and informality: Evidence from Brazil, manuscript. Robertson, R. 2005. Relative prices and wage inequality: Evidence from Mexico, in Journal of International Economics, Vol. 64, No. 2, pp. 387-409. Satchi, M.; Temple, J. 2009. Labor markets and productivity in developing countries, in Review of Economic Dynamics, Vol. 26, pp. 333-345. Slonimczyk, F. 2012. The eect of taxation on informal employment: Evidence from the Russian at tax reform, in  Informal Employment in Emerging and Transition Economies, Vol. 34 of Research in Labor Economics, chapter 2, pp. 55-99, Lehmann, H. and Tatsiramos, K. (eds.). 33 Verhoogen, E. A. 2008. Trade, quality upgrading and wage inequality in the Mexican manufacturing sector, in Quarterly Journal of Economics, Vol. 123, No. 2, pp. 489- 530. Waseem, M. 2018. Taxes, informality and income shifting: Evidence from a recent Pak- istani tax reform, in Journal of Public Economics, Vol. 157(C), pp. 41-77. Zenou, Y. 2008.  Job search and mobility in developing countries: Theory and policy implications, in Journal of Development Economics, Vol. 86, pp. 336-355. Zhang, Y.; Wan, G. 2006. Globalization and the urban poor in China, United Nations University World Institute for Development Economics Research (WIDER) Paper No. 2006/42. 34 A Dynare equations To solve the model, we use the Dynare software (see Adjemian et al. (2011)). After solving the steady state of the model, we use the following equation set in order to obtain the equilibrium paths. • The equilibrium price index 1−φ 1−θ ∗ 1−φ ˜d,t 1=ρ 1−θ Nd,t ˜∗ +ρ 1−θ 1−θ x,t Nx,t • Average export productivity θ −φ θ−1 z˜x,t ˜− ρ θ 1−θ x,t Nx,t Yt C∗ = fx,t kp − (θ − 1) τt • Labor market clearing ˜d,t y ˜x,t y ZF lF t + ZI lIt = Nd,t + Nx,t τt + Ne,t fe,t + Nx,t fx,t z˜d ˜x,t z • Law of motion of employment ljt = (1 − λj )ljt−1 + qjt−1 Vjt−1 • New variety (product) creation fe,t+1 Nx,t+1 fx,t+1 Ct−γc c +1 1 + τt ρ˜d,t+1 fe,t −Nd,t+1 fe,t 1 = (1 − δ )βEt −γc ˜ +1 y c Ct 1 + τt+1 ρ ˜d,t 1 + (θ−1)fe,t d,t z ˜d +N x,t+1 τt+1 y ˜ ˜x,t+1 x,t+1 Nd,t+1 z • Job creation Ct− γc +1 1 + τt c qjt qjt f 1 = βEt −γc c (1 − λj ) + ϕt+1 Zjt+1 − wjt+1 (1 + τjt+1 ) Ct 1 + τt+1 q jt+1 κ j • Wage determination η b e1+ Ft e1+ It wF t = w − ϑ − ϑ 1 − τF t (1 + τtc ) 1+ 1+ f w 1−η κF 1 + τF t 1 − τF t+1 + f ϕt ZF t + + (1 − λF ) − (1 − λF − ιF t ) f w 1 + τF t qF t 1 + τF t+1 1 − τF t b e1+ Ft e1+ ιIt wIt = η c −ϑ − ϑ It + (1 − η ) ϕt ZIt + κI (1 + τt ) 1+ 1+ qIt 35 • Search intensity 1−η VF t 1 − τtw +1 κF ϑeF t = χF η eF t Ut (1 + τtf+1 )(1 + τtc+1 ) qF t 1−η VIt 1 κI ϑeIt = χI η eIt Ut 1 + τtc+1 qIt • Euler equation for domestic bond holding −γc c Cat+1 1 + τt (1 + ψat+1 ) = (1 + it+1 )βEt −γc Cat 1 + τtc+1 • Euler equation for foreign bond holding −γc c Cat+1 Qt+1 1 + τt (1 + ψa∗ t+1 ) = (1 + i∗ t+1 )βEt −γc Cat Qt 1 + τtc+1 • Bond market clearing at+1 + a∗ t+1 = 0 • Net foreign assets (at+1 − at ) + Qt (a∗ ∗ ∗ ∗ t+1 − at ) = it at + Qt it at + Qt Nx,t ρ ˜x,t y ∗ ∗ ˜x,t − Nx,t ρ ˜x,t y ∗ ˜x,t • Law of motion for the stock of producing plants Nd,t+1 = (1 − δ )(Nd,t + Ne,t ) • Share of exporting plants kp zmin kp Nx,t = α θ−1 Nd,t ˜x,t z • Export productivity cuto 1 ˜x,t = α θ−1 zx,t z • Average price of a domestic variety φ ϕt ˜d,t = ρ φ−1z ˜d • Average price of an exported variety φ τt ϕt ˜x,t = ρ φ − 1 Qt z ˜x,t 36 • Average output of a domestic variety θ −φ y ˜− ˜d,t = ρ φ 1−θ d,t Nd,t Yt C • Average output of an exported variety θ −φ y ˜− ˜x,t = ρ φ 1−θ x,t Nx,t Yt C∗ • Aggregate demand YtC = ωCat + (1 − ω )(CF t + CIt + Cut ) + κF VF t + κI VIt + Gt • Formal workers' consumption 1 − τtw CF t = wF t lF t 1 + τtc • Unemployed agents' consumption bt (1 − lt ) Cut = 1 + τtc • Informal workers' consumption lIt CIt = wIt 1 + τtc • Total consumption Ct = ωCat + (1 − ω )(CF t + CIt + Cut ) • Unemployment Ut = 1 − lF t − lIt • Job lling rate 1−ε ejt Ut qjt = χj Vjt • Job nding rate ε Vjt ιjt = χj ejt ejt Ut • Productivity shock ∗ log Zjt = φZ 1 log Zjt−1 + φZ 2 log Zjt−1 + Z jt • Government spending Gt = τtc [ωCat + (1 − ω )(CF t + CIt + Cut )] + (τtw + τtf ) wF t lF t − bt Ut 37 B Steady states Figure 4: Wage inequality Wage inequality-H Wage inequality-F 18 72 16 70 69.2 14.4 14 68 12 66 65.8 65.7 10.7 10.6 10 64 8 62 6 60 Initial TL TL&TR Initial TL TL&TR Note: H and F indicate respectively the Home country (i.e. the developed country) and the Foreign country (i.e. the emerging country). Wage inequality is dened by wF t /wIt − 1. 38 Figure 5: Goods markets Domestic producers−H Exporters−H Export cutoff−H 22 1.5 6 21 20 4 19 1 Initial tl tr Initial tl tr Initial tl tr Price in H−H Price in F−H Real exch. rate 3.2 3.345 1.01 3.1 3.34 1.005 3 3.335 1 Initial tl tr Initial tl tr Initial tl tr Domestic producers−F Exporters−F Export cutoff−F 19 5 1.5 18 4 17 3 1 Initial tl tr Initial tl tr Initial tl tr Price in H−F Price in F−F Share of exporters 3.2 3.16 0.3 3 3.155 0.2 2.8 3.15 Initial tl tr Initial tl tr Initial tl tr Note: H and F indicate respectively the Home country (i.e. the developed country) and the Foreign country (i.e. the emerging country). 39 Figure 6: Labor markets Formal wage−H Formal employment−H Share of formal emp−H 0.8725 0.928 4.9 0.872 4.8 0.927 4.7 0.8715 Initial tl tr Initial tl tr Initial tl tr Informal wage−H Informal employment−H Share of informal emp−H 0.069 2.9 0.068 0.0725 2.85 0.067 0.072 Initial tl tr Initial tl tr Initial tl tr Formal wage−F Formal employment−F Share of formal emp−F 0.621 3.8 0.682 0.62 0.619 0.681 3.6 0.618 0.617 0.68 Initial tl tr Initial tl tr Initial tl tr Informal wage−F Informal employment−F Share of informal emp−F 3.35 0.2905 0.29 0.3185 3.3 0.2895 0.289 3.25 0.2885 0.318 Initial tl tr Initial tl tr Initial tl tr 40 Figure 7: Unemployment Unemployment−H Formal search effort−H Informal search effort−H 1.04 1.385 1.035 0.06 1.38 1.03 0.0598 1.375 1.025 0.0596 1.37 1.02 0.0594 1.365 1.015 1.36 1.01 Initial tl tr Initial tl tr Initial tl tr Unemployment−F Formal search effort−F Informal search effort−F 0.092 1.265 1.22 0.0915 1.215 1.26 0.091 1.21 1.255 0.0905 1.205 0.09 1.25 1.2 Initial tl tr Initial tl tr Initial tl tr Note: H and F indicate respectively the Home country (i.e. the developed country) and the Foreign country (i.e. the emerging country). 41