外人直接投資（foreign direct investment, FDI）在經濟理論中是相當熱門的議題，它代表了讓地主國（host country）國資本累積、技術進步在短時間內快速增加的可能，因此許多國家往往會採取某些吸引外資的政策，再搭配國內制度或是貿易政策的改變，以追求經濟成長。然而，外來直接投資對地主國可能產生的所得重分配的影響，本文即是對此做一深入探討，並以台灣製造業資料來研究外來直接投資是否會擴大工資不均等的情形。 研究期間從1981～2004年共24年，依產業特性將製造業分為十大類，分別採取兩種不同的迴歸模型，包括自我迴歸落遲分配模型（auto regressive distributed lag model, ARDL model）與縱橫資料（panel data）迴歸模型等。實證模型上由生產理論出發，選擇作為解釋工資不均等的變數包括外人直接投資比例、出口比例、進口比例及產出成長率等四個變數。由實證結果得到以下結論： （1）就個別產業來看，FDI對台灣製造業工資不均等的影響並不一致，反應出產業特性不同，FDI所扮演的角色也不盡相同。其中FDI會惡化皮革與毛皮製造業的工資不均等情形，減輕橡膠及塑膠製品製造業與非金屬製品製造業的工資不均等情形，對其他製造業則是無明顯影響。 （2）就整體製造業的情形來看， FDI對工資的不均等的淨效果為正，但效果不大；出口、產出成長率有輕微使工資不均等擴大的情形，而進口則是可輕微縮減工資不均等的狀況。 （3）若是將十大製造業依產品特性區分為「民生」、「化學」、「機械」電子等三大工業，則可以發現FDI對民生工業有明顯擴大工資不均等的情形，在其他兩大工業則是無顯著影響。

La acumulación de Capital Humano es fundamental para explicar recientes procesos de crecimiento económico experimentados en numerosos países, los cambios en la distribución de la renta salarial, en la localización de la actividad productiva, del comercio, etc... Por ello, consideramos fundamental avanzar en el conocimiento que dicha acumulación haya podido tener en una economía como la española, caracterizada recientemente por un crecimiento intenso en la educación de los trabajadores. Por este motivo se realizan tres análisis. En el primero, descomponemos la distribución de los salarios para conocer qué factores han afectado a su cambio desde 1980. Dedicamos especial atención al efecto que el cambio educativo haya podido tener en la desigualdad salarial. En un segundo análisis descomponemos el cambio en la brecha salarial entre trabajo cualificado y no cualificado entre cambios en la demanda y oferta de cualificación. Por último, estimamos las externalidades del capital humano en España a nivel regional. / Human Capital accumulation is crucial for explaining countries' recent economic growth episodes, changes in income and wage distribution, production localization, trade, etc. Thus, we consider important enhance the knowledge for the Human Capital accumulation effects on Spanish economy, characterized by intensive growth of educated workers. Then, we do three analysis. First, we decompose the wage distribution to know which factors are behind its changes since 1980. We give special attention to education effects on that wage inequality change. In our second analysis, we decompose the more educated wage premium change between changes in demand and supply of skills. Finally, our third analysis try to estimate human capital externalities for Spanish regions.

[pt] Esta dissertação tem como objetivo decompor a desigualdade de rendimentos entre raças e gêneros, em montantes explicáveis pelas escolhas educacional e ocupacional e pela discriminação. Foi utilizada uma metodologia micro- econométrica de simulações contra-factuais, conforme Bourguignon, Ferreira e Lustig (1998), com os dados da PNAD de 1996, incluindo no modelo rendimento mensal, educação, posição na ocupação, entre outras variáveis. Através deste procedimento conseguiu-se atribuir um alto poder explicativo do termo de discriminação, principalmente no caso das mulheres, em que foi computado mais de 100%. Para entender a diferença entre negros e brancos também pode-se recorrer, parcialmente, à desigualdade nas escolhas educacionais, e menos à desigualdade ocupacional. Tendo computado simultaneamente a desigualdade por raça e gênero, nota-se o efeito do gênero sobrepor-se, implicando na perda de poder explicativo da ocupação para o termo de discriminação. Ressalta-se que para homens e mulheres negros a educação dos pais pode ser um importante determinante da desigualdade nos rendimentos. Conclui-se que as políticas sociais voltadas para o combate das desigualdades de rendimentos devem focalizar-se na educação e na transmissão intergeracional, além de procurar medidas que minimizem a diferenciação na remuneração das características. Recomenda-se ainda uma análise temporal, bem como um estudo mais aprofundado sobre os efeitos da educação dos pais. / [en] This thesis is concerned with the analysis of the inequality of earn between races and genders in amounts explained by educational and occupational choices among other characteristics. A micro econometric counterfactual simulation methodology, following Bourguignon, Ferreira e Lustig (1998) was used. The data basis was the PNAD 1996, including monthly earn, education, position, among others. Through this procedure, it was possible to ascertain the high accountability of the discrimination term, mainly referred to women, where over than 100 percent was found. In order to understand the inequality between black and white populations the explanation can be partially referred to the inequality of their educational choices, and also to their occupational choices. For black males and females parents education seems to be a important earn inequality determinant. As a conclusion, social policies concerned with the earn inequalities reduction must aim at education and at the intergenerational transmission of education inequality among blacks, and search for measures that will reduce the differentiation of compensation characteristics. A time series analysis is suggested in order to access a possible decreasing tendency in the term of discrimination, as well as a profound study of the parental education effect on race inequality.

nÃo hà / Teorias em Economia Internacional sÃo desenvolvidas e testadas empiricamente, esta tese pretende contribuir com algumas delas. Inicialmente, dever-se-à abordar a chamada teoria de Cones de DiversificaÃÃo. Para tanto, o primeiro capÃtulo propÃe-se a separar em dois cones os estados do Brasil que apresentam semelhanÃas nas dotaÃÃes de fatores. Isto serà realizado atravÃs de um modelo economÃtrico SUR para dezoito indÃstrias de transformaÃÃo e duas amostras anuais, em 1997 e 2007. DiferenÃas salariais entre os cones tambÃm serÃo medidas com a mesma metodologia. A teoria de cones de diversificaÃÃo à revalidada uma vez que existe diferenciaÃÃo horizontal entre os produtos. O segundo capÃtulo elabora um modelo economÃtrico que delineia os efeitos da abertura comercial sobre os salÃrios dos trabalhadores pouco e altamente qualificados. Utilizou-se o arcabouÃo teÃrico da equaÃÃo minceriana para incorporar caracterÃsticas dos trabalhadores. Este trabalho objetiva analisar se, dada maior abertura comercial, bem como outros fatores relevantes, tais quais; experiÃncia, anos de estudo, intensidade tecnolÃgica do setor ao qual se emprega e quantidade de horas trabalhadas, houve aumento da renda do trabalhador industrial dos estados do Nordeste do Brasil para os anos da amostra. Adicionalmente, uma dummy à inserida no modelo com o objetivo de diferenciar os trabalhadores empregados nos setores de alta e baixa tecnologia. Encontram-se evidÃncias de que volume de exportaÃÃes, horas de trabalho, anos de estudo e experiÃncia impactaram positivamente no retorno do trabalhador. Maior abertura comercial apresenta grande influÃncia positiva sobre os salÃrios dos trabalhadores das indÃstrias analisadas. O Ãltimo capÃtulo apresenta um modelo gravitacional que serà aplicado a fim de estudar os determinantes do fluxo comercial entre o Estado do Cearà e os demais estados brasileiros. Este capÃtulo dirigiu sua atenÃÃo aos fatores que influenciam o fluxo de comÃrcio do Estado do Cearà para os demais estados do Brasil. Entre estes fatores, apontou-se inicialmente o espaÃo que separa os centros produtivos como um fator que atua influenciando negativamente o comÃrcio. Ainda, analisou-se o impacto das variÃveis Produto Interno Bruto e tamanho populacional, encontra-se que ambas exercem impacto positivo para o fluxo comercial. / Theories in International Trade are developed and empirically tested. This thesis aims to contribute with some of them. Initially, it will address the so-called Theory of Diversification Cones. Thus, the first chapter proposes to separate Brazilian states into two cones which have similarities in factor endowments. This will be accomplished through an SUR econometric model with eighteen manufacturing industries and two annual samples, in 1997 and 2007. Wage differences between the cones will also be measured with the same methodology. The Theory of Diversifiation Cones is renewed since there is horizontal differentiation between products. The second chapter develops an econometric model that outlines the effects of trade liberalization on wages of low and high-skilled workers. We used the theoretical framework of the Mincerian equation to incorporate characteristics of workers. This study aims to examine whether, given greater trade openness as well as other relevant factors, such that; experience, years of education, technological intensity of the sector to which it is used and the amount of hours worked, an increase in income of the industrial worker in the states of Northeast Brazil for the years of the sample. In addition, a dummy is inserted into the model with the aim to differentiate the workers employed in low and high-tech industries. There are evidences that export volume, hours, years of study and experience have a positive impact on the worker's return. Greater trade openness has a large positive influence on the wages of workers in industries analyzed. The final chapter presents a gravity model which will be applied for the purpose of studying the determinants of trade flows between the State of Cearà and other Brazilian states. This chapter turned its attention to the factors that influence the trade flow in the State of Cearà in other states of Brazil. Among these factors, the separation between the production centers was initially pointed as a factor that acts negatively influencing trade. The impact of the variables GDP and population size was also analyzed, and it was expected that both of them exert positive impact on trade flows.

Submitted by Amanda Cappellazzo Arabage (amanda_arabage@outlook.com) on 2018-06-11T12:56:15Z No. of bitstreams: 1 Tese Amanda C A.pdf: 2902865 bytes, checksum: d10ee483dca9cc724f63a79e7e0f8643 (MD5) / Approved for entry into archive by Katia Menezes de Souza (katia.menezes@fgv.br) on 2018-06-11T13:23:23Z (GMT) No. of bitstreams: 1 Tese Amanda C A.pdf: 2902865 bytes, checksum: d10ee483dca9cc724f63a79e7e0f8643 (MD5) / Approved for entry into archive by Isabele Garcia (isabele.garcia@fgv.br) on 2018-06-12T19:05:57Z (GMT) No. of bitstreams: 1 Tese Amanda C A.pdf: 2902865 bytes, checksum: d10ee483dca9cc724f63a79e7e0f8643 (MD5) / Made available in DSpace on 2018-06-12T19:05:57Z (GMT). No. of bitstreams: 1 Tese Amanda C A.pdf: 2902865 bytes, checksum: d10ee483dca9cc724f63a79e7e0f8643 (MD5) Previous issue date: 2018-05-21 / This thesis is composed of three essays in applied microeconomics. In all essays, we investigate different aspects related to the Brazilian formal labor market using data from the Relação Anual de Informações Sociais - RAIS. In the first two chapters, we examine the impacts of two educational policies on labor market outcomes of those involved. In the third chapter, we study wage inequality in the formal labor market. We aim to contribute to the debate on inequality reduction policies by providing additional empirical evidence for Brazil. In the first chapter, we investigate the impacts of the provision of free child care for children aged 0-3 on mothers and fathers’ labor market outcomes in the formal sector of the economy. We exploit an exogenous variation introduced by the fact that, for the 2008 school year, access to public day care centers was randomly assigned in the municipality of Rio de Janeiro. Our results suggest that the provision of this service had no impacts on parents’ probabilities of being employed in the formal sector and related labor market outcomes (e.g. hours worked, wages) in the short or medium terms. We argue that high labor market participation, relatively low flexibility of the formal labor market and the fact that the majority of children eventually enroll in a public day care center or school in the years following the intervention may be possible explanations for these results. In the second chapter, we investigate if the introduction of quotas in public universities affects employers’ perception of its students and graduates, regardless of their actual quota status. We explore the different timing of introduction of quota systems in two public universities in the State of Rio de Janeiro to estimate impacts of this policy on its students and graduates’ probabilities of being employed in the formal sector and hourly wages for those employed. Our results suggest that the introduction of quotas does not seem to have effects on the probability of being employed in the formal sector but there seems to be some negative effects on hourly wages for those who are employed in the formal sector for some of the cases considered. In the third chapter, we study the decrease in wage inequality among working age men in the Brazilian formal labor market from 1994 to 2016. We decompose inequality into its permanent and transitory components and find that the downward trend in wage inequality can be attributed to reductions in the transitory component from 1994 to 2005 (related to the economic stabilization process), and in the permanent component from 2010 to 2016. Our results also suggest that the permanent component has an important role in explaining wage inequality in Brazil, with age and education being accountable for a great share of permanent inequality. / Esta tese é composta por três artigos na área de microeconomia aplicada. Em todos os artigos, são investigados aspectos relacionados ao mercado de trabalho formal brasileiro, utilizando dados da Relação Anual de Informações Sociais - RAIS. Nos dois primeiros capítulos, investigam-se os impactos de duas políticas educacionais em resultados no mercado de trabalho dos indivíduos envolvidos. No terceiro capítulo, estuda-se a desigualdade de salários no mercado de trabalho formal. Com os artigos, pretende-se contribuir para o debate sobre políticas de redução da desigualdade, fornecendo evidências empíricas adicionais para o Brasil. No primeiro capítulo, são investigados os impactos do oferecimento de creches gratuitas para crianças de 0 a 3 anos sobre os resultados de suas mães e pais no mercado de trabalho formal. Para isso, utiliza-se a variação exógena introduzida pelo fato de que, para o ano letivo de 2008, o acesso a vagas em creches públicas no município do Rio de Janeiro foi definido a partir de um processo de aleatorização. Os resultados sugerem que este serviço não teve impactos sobre as probabilidades das mães e pais de estarem empregados no setor formal e sobre outros resultados relacionados (por exemplo, horas trabalhadas e salários) tanto no curto quanto no médio prazos. Argumenta-se que a alta participação no mercado de trabalho, a reduzida flexibilidade do mercado de trabalho formal e o fato de que a maioria das crianças eventualmente foi matriculada em uma creche ou escola pública nos anos seguintes à intervenção são possíveis explicações para esses resultados. No segundo capítulo, é analisado se a introdução de cotas em universidades públicas afeta a percepção dos empregadores sobre seus alunos e ex-alunos, independentemente de eles terem ingressado no ensino superior via cotas. Para isso, utiliza-se o fato de que a adoção de sistemas de cotas em duas universidades públicas do Estado do Rio de Janeiro ocorreu em momentos distintos. Dessa forma, é possível estimar os impactos dessa política na probabilidade de seus alunos e ex-alunos de estarem empregados no setor formal e nos salários por hora para aqueles que estão empregados. Os resultados sugerem que a introdução de cotas não parece ter tido efeitos sobre a probabilidade de estar empregado no setor formal, mas são encontrados alguns efeitos negativos sobre os salários para aqueles que estão empregados nesse setor em alguns dos casos considerados. No terceiro capítulo, analisa-se a redução da desigualdade de salários entre homens em idade ativa no mercado de trabalho formal brasileiro de 1994 a 2016. A desigualdade é decomposta em seus componentes permanente e transitório e verifica-se que a tendência de queda da desigualdade salarial pode ser atribuída a reduções no componente transitório de 1994 a 2005 (relacionadas ao processo de estabilização econômica), e no componente permanente de 2010 a 2016. Os resultados também sugerem que o componente permanente tem um papel importante na desigualdade salarial no Brasil, sendo idade e educação responsáveis por uma parcela expressiva da desigualdade permanente.

Subsidising investment in lagging regions is an important regional policy instrument in many countries. Some argue that this instrument is not specific enough to concentrate the aid towards the regions that are lagging behind most, because investment subsidies benefit capital owners who might reside elsewhere, possibly in very rich places. Checking under which conditions this is true is thus highly policy relevant. The present paper studies regional investment subsidies in a multiregional neoclassical dynamic framework. We set up a model with trade in heterogeneous goods, with a perfectly integrated financial capital market and sluggish adjustment of regional capital stocks. Consumers and investors act under perfect foresight. We derive the equilibrium system, show how to solve it, and simulate actual European regional subsidies in computational applications. We find that the size of the welfare gains depends on the portfolio distribution held by the households. If households own diversified asset portfolios, we find that the supported regions gain roughly the amounts that are allocated to them in the form of investment subsidies. If they only own local capital stocks, a part of the money is lost through the drop in share prices. From the point of view of total welfare, the subsidy is not efficient. It can lead to a welfare loss for the EU as a whole and definitely leads to welfare losses in the rest of the world, from where investment ows to the supported EU regions.

Cette thèse est composée de trois essais; dans le premier essai je traite les questions de la détermination, de la variance et des répercussions de la corruption. J’ai montré que la corruption est déterminée par la part des travailleurs non qualifiés sur la population. Si cette part est large alors il existe une corruption, si elle est faible la corruption est inexistante, et pour des niveaux intermédiaires, il existe une multiplicité d’équilibres. La corruption augmente les inégalités salariales entre travailleurs qualifiés et non qualifiés, et une perte de bien-être. Dans le deuxième essai je traite la question de lutte contre la corruption à travers l’incitation salariale. Avec une technologie de contrôle endogène, je montre que le gouvernement peut mieux accepter la corruption lorsqu’il est coûteux de contrôler. Lorsqu’il est optimal de combattre alors le gouvernement peut le faire soit à travers des salaires d’efficience ou soit par le contrôle. Néanmoins le rôle des salaires d’efficience dans la lutte contre la corruption est moindre dans les sociétés avec un niveau de malhonnêteté élevé. Le troisième essai traite la malédiction des ressources naturelles. Je montre que l'éducation et la corruption sont déterminées conjointement ; les ressources naturelles affectent l’incitation à investir en éducation et en ‘rent-seeking’ ce qui en retour affecte la croissance. En outre, la relation entre une abondance et la malédiction des ressources naturelles n’est pas monotone. Pour un niveau d’inégalité d’accès à l’éducation faible et un coût élevé de participation dans la vie politique, un niveau de croissance élevé et la trappe à la pauvreté coexistent. / This dissertation consists of three essays. The topics cover determination, variance and repercussions of corruption (essay one), corruption deterrence through wage incentives (essay two), and natural resource curse (essay three). In the first essay, I show that for a larger population of unskilled labor, there is a widespread corruption and for a smaller population there is no corruption. For the intermediate levels there are multiple equilibria. On its consequences, corruption increases wage inequality between skilled and unskilled workers, and results in output and welfare losses. In the second essay, I argue that deterring corruption through efficiency wage may become prohibitively expensive. With endogenous monitoring technology that allows capturing the dual role of auditing, as a complement with and as a substitute for wage incentives, I find that the government is better-off accepting corruption when it is costly to monitor. When it is optimal to deter bribery, the government can do it either through efficiency wages or monitoring. The role of efficiency wages decreases in societies with higher level of dishonesty. In the third essay, I build a theory explaining a resource curse. In contrast to the existing literature which generally considers low education, corruption and natural resources separately, I combine three strands of literature. Natural resources affect incentives to invest in education and rent seeking that in turn affects growth. Second, the relationship between resource-abundance and resource-curse is non-monotonic. For low inequality in access to education and high cost of political participation, high-growth and poverty-trap equilibria co-exist.

Las tecnologías de la información y la comunicación constituyen hoy en día una herramienta fundamental para el trabajo humano en el futuro, pues ha contribuido a reducir las distancias y ha alcanzado la información a muchas personas al mismo tiempo. Además, la automatización ha mejorado notablemente las perspectivas del trabajo industrial aminorando drásticamente el trabajo humano. Para ello, la inteligencia artificial ocupa un lugar trascendental, pues permite desarrollar la automatización a un nivel más alto, pues facilita la actividad humana a través de las comunicaciones y la ingeniería, entre otros. Para desarrollar la presente propuesta, se expone, en primer lugar, un contexto tecnológico en el cual se revisarán las diferentes variables encontradas como por ejemplo automatización, industria 4.0, inteligencia artificial y tecnologías de la información. En segundo lugar, a través de un contexto socio cultural, se revisan variables como adaptación, cultura, brechas políticas e inestabilidad laboral. En tercer lugar, se presentarán y contrastarán las principales posturas acerca de los efectos a largo plazo de las nuevas tecnologías de la información y la automatización en el trabajo humano con la finalidad de presentar la controversia sobre los efectos a largo plazo de las nuevas tecnologías de la información y automatización en el trabajo humano. En conclusión, es necesario destacar el desarrollo de la inteligencia artificial y su impacto en las tecnologías de la información, el entorno asociado a las personas y su entendimiento de la idiosincrasia, así como la información y la automatización en el trabajo humano. / Information and communication technologies are today a fundamental tool for human work in the future, as it has contributed to reducing distances and has reached information to many people at the same time. In addition, automation has markedly improved the prospects for industrial work by drastically reducing human labor. For this, artificial intelligence occupies a transcendental place, since it allows to develop automation at a higher level, since it facilitates human activity through communications and engineering, among others. To develop this proposal, a technological context is exposed, first of all, in which the different variables found will be reviewed, such as automation, industry 4.0, artificial intelligence and information technologies. Second, through a socio-cultural context, we will review variables such as adaptation, culture, political gaps, and job instability. Third, the main positions on the long-term effects of new information technologies and automation on human work will be presented and contrasted with the aim of presenting the controversy on the long-term effects of new information technologies. information and automation in human work. In conclusion, it is necessary to highlight the development of artificial intelligence and its impact on information technologies, the environment associated with people and their understanding of idiosyncrasy, as well as information and automation in human work. / Trabajo de Suficiencia Profesional

Diese Dissertation umfasst vier Aufsätze welche die Effekte von nicht-zufälliger Selektion betrachten. Im ersten Aufsatz wird analysiert welche Rolle die geänderte Zusammensetzung der Erwerbsbevölkerung für die Entwicklung der Lohnungleichheit spielt. Vollzeit-Erwerbstätige weisen zunehmend Episoden von Teilzeitarbeit und Erwerbsunterbrechungen in ihren Biographien auf. Dies hat maßgeblich zum Anstieg der Lohnungleichheit beigetragen. Der zweite Aufsatz betrachtet die Effekte von gesunkener Arbeitslosigkeit auf Lohnungleichheit.. Wenn sich die Selektion in Vollzeit oder Arbeitslosigkeit ändert, kann dies zu steigender Lohnungleichheit führen. Dies ist insbesondere dann der Fall wenn Personen neu in Beschäftigung kommen, die eine Negativauswahl der Erwerbsbevölkerung sind. Jedoch zeigt sich, als Resultat einer Analyse mit selektionskorrigierten Quantilsregressionen, dass diese veränderte Selektion nicht zum Anstieg der Lohnungleichheit beigetragen hat. Im dritten Aufsatz geht es um nicht-monotone Selektion bei Regression Discontinuity Designs (RDD). Die Annahmen, auf welchen RDDs beruhen, können verletzt sein wenn sich sowohl Individuen in die Maßnahmengruppe hinein, als auch aus ihr heraus selektieren. Der Aufsatz beschreibt diesen Selektionsmechanismus und zeigt seine praktische Relevanz. Zudem wird ein Spezifikationstest vorgestellt um das Problem im Vorfeld der Analyse zu erkennen. Der vierte Aufsatz untersucht Wählerpräferenzen für das Geschlecht politischer Kandidaten. Es wird analysiert ob das Zusammenspiel von Berufsinformationen und Geschlecht die geringere Repräsentation von Frauen erklären kann. Dazu wurde ein Feldexperiment durchgeführt. Es zeigt sich dass weibliche Kandidaten einen Stimmvorteil genießen solange keine Berufsinformationen angegeben sind. Sobald jedoch die Berufe der Kandidaten bekannt sind kehrt sich dies in einen Stimmvorteil für männliche Kandidaten um. / This thesis comprises four essays which study effects of non-random selection. The first essay analyzes the role of changes in labor force composition for the development of wage inequality among full-time workers. Of particular interest are the effects of increasingly common episodes of temporary part-time work and nonemployment among full-time workers. Such episodes, have contributed substantially to the rise in wage inequality. The second essay studies the effects of declining unemployment on inequality of wages. Changing selection over time between unemployment and full-time work could lead to increasingly diverse wages, particularly if sinking unemployment implies an influx of negatively selected workers into employment. However, results from a selection corrected quantile regression approach show that changing selection with respect to unobservables is not a contributor to the rise in wage inequality. The third essay studies non-monotonic selection in regression discontinuity designs (RDD). When similar numbers of individual select into and out of treatment simultaneously, the identifying assumption of the RDD can be violated. The essay describes the selection mechanisms and demonstrates it’s practical relevance. It then suggests an enhancement to the standard specification tests for RDDs, which can detect non-monotonic sorting in advance. The fourth essay studies voter’s valuation of candidate gender. It examines whether the presence of profession information coupled with voter preferences for stereotypical male occupations may explain part of the gender gap in parliaments. The analysis is conducted as a field experiment built into an exit-poll of voters in Germany in 2014. The results show a vote share bonus for women in the absence of profession information. Once voters know the profession of candidates, however, this changes towards a small edge for male candidates.

We set up a trade model with heterogeneous firms and a worker population that is heterogeneous in two dimensions: workers are either skilled or unskilled, and within each skill category there is a continuum of abilities. Workers with high abilities, both skilled and unskilled, are matched to firms with high productivities, and this leads to wage differentials within each skill category across firms. Self-selection of the most productive firms into exporting generates an exporter wage premium, and our framework with skilled and unskilled workers allows us to decompose this premium into its skill-specific components. We employ linked employer-employee data from Germany to structurally estimate the parameters of the model. Using these parameter estimates, we compute an average exporter wage premium of 5 percent. The decomposition by skill turns out to be quantitatively highly relevant, with exporting firms paying no wage premium at all to their unskilled workers, while the premium for skilled workers is 12 percent.