Final Report Summary - TRADE AS AID (Trade Preferences as Development Aid: Theory and Evidence)
The most successful development stories of the past half a century have been those of Japan, South Korea, Hong Kong, and more recently mainland China. All of these “growth miracles” were strongly export-oriented, with all countries continuously upgrading and diversifying their exports by adopting the more advanced technologies of abroad and investing in both physical and human capital. This fact that has led to the increased use of trade policy by developed countries as a development tool to encourage growth through exporting in developing countries. In particular, in 2000, the African Growth and Opportunities Act (AGOA) granted a set of least developed countries (LDC) special access to the U.S. markets for apparels and other products without the strict content requirements on locally produced inputs that had usually accompanied such programs. Although, these preferences are politically very popular, we have little understanding of how well they work, both in terms of transferring wealth and encouraging growth. The current state of the art of international economics does not currently have a theory for how to best design them nor whether they are superior to other forms of aid such as direct aid. Improving on this is the goal of this project.
The three main research objectives of the project are i) to properly model and examine the efficiency of trade policy as a tool for transferring wealth to a beneficiary country, ii) to test the empirical predictions using a rich data set, and iii) to understand how trade policy and integration into the international economy in general can encourage economic growth in the least developed countries.
Part i) exists in complete working-paper format and will shortly be submitted. I have developed a general equilibrium model, which allows for normative assessment of trade policy as a wealth transfer. The main argument of the original grant application is proven formally (deviations from a nationally optimal trade policy will have only second order effects on nation welfare but generally positive first order effects on foreign welfare and is therefore very generally an efficient way of transferring wealth). The largest deviation from the original grant proposal is in scope where the final model includes political economy considerations, internal equality and multiple countries. This enriches the scope and makes the analysis more broadly applicable. All results have been demonstrated, but the exposition is still incomplete.
Part ii) is still incomplete but a working paper will be submitted in the fall. We have solved some empirical problems and converged on a final empirical strategy and find that the data is broadly consistent with the predictions of our model.
Part ii) partly uses Gravity Equations. In a related paper Pankaj Ghemawat and I show that randomness in productivity draws for firms can accounts for part of the variation in trade flows across countries. The basic argument is as follows: from the perspective of a country entrepreneurial quality is a random draw. A large country, such as the United States, will have many “draws” an as a consequence will have many successful companies spread across several industries. A smaller country, say, Korea, might only have one really successful company that ends up dominating. As a consequence, smaller countries will have more “noisy” distributions of exports across product categories. We have done the formal and empirical analysis and have an early version of a working paper.
Part iii) has progressed substantially more and has in fact given rise to a number of additional project. As the general equilibrium model of part i) is static and takes technology as a given, it can treat the wealth transfers through trade policy but is silent on the ability of trade policy as a means of transferring technology. Such questions are best answered in a model of endogenous technology and we have been developing exactly such a framework. It turns out that the prospective for technological transfers is deeply connected to the progress of technological development in the Western world: Suppose, technological development increasingly substitutes for instead of complementing low-skill labor (Acemoglu and Autor, Handbook of labor economics, 2013) such that the need for low-skill labor in modern production facilities is reduced. When the need for low-skill labor is lower the incentive to move production to low-labor cost countries is lower and the mechanism through which South Korea, China and Japan grew might be weaker in the immediate future than in the second half of the twentieth century. Needless, to say these mechanism will be essential for understanding the ability of trade policy to encourage growth. The intend of the original grant proposal was to extend one of the existing models of endogenous growth to this context. However, we argue that the present theoretical foundations are insufficient to address these issues and instead, along with David Hemous (University of Zurich), I have developed an endogenous growth model of directed technological change which directly addresses the question of when technology might be labor-replacing. This paper is complete and is under review at the American Economic Journal: Macroeconomics.
I have started an additional project (joint with David Hemous and Antoine Dechezlepretre at the London School of Economics) as an empirical investigation of the predictions in the theoretical framework mentioned above. In particular, we use patent codes to classify the universe of patents as either “automation patents” or “non-automation patents” and we ask how the wages facing firms affect these firms' tendency to focus on innovation in automation technology. There is not a yet a working paper, but the preliminary results suggests that 1 per cent higher wages for the low-skill workers employed raises the patenting of automation patents by around 1.1 per cent.
In addition, the progress of technological change is determined by a number of factors specific to the local economy. In a paper with David Hemous, I show that the extend to which firms are engaged in long-term relationships – as a means of overcoming incomplete contract enforcement – determines the direction of technological change. When engaged in strong relationships firms are more likely to focus on relationship specific innovation, whereas economies with less strong business relationships are more prone to larger more general innovations. Using patent data we show that this tendency is borne out by the data. This paper had a Revise and resubmit request from the Journal of Europen Economic Association and has been resubmitted.
From the work on technological change and income inequality sprung a paper on the spill-over effects of income inequality: Suppose technological change or deregulation increases income inequality in a particular occupation, say, financial managers. Then this increase in income inequality will spill over into other occupations such as physicians as the wealthier financial managers compete for the best doctors. We use data from the federal census in the United States to show that this can explain a substantial part of the increase in income inequality. This paper is basically done, but the analysis was carried out on censored Federal Census data. We are presently waiting for the uncensored data. This paper is with David Hemous and Jeffrey Clemens of the University of San Diego.
The expected final result of the project is to assess the potential of trade policy as a means of development aid, and in particular the optimal design of such policies.
The three main research objectives of the project are i) to properly model and examine the efficiency of trade policy as a tool for transferring wealth to a beneficiary country, ii) to test the empirical predictions using a rich data set, and iii) to understand how trade policy and integration into the international economy in general can encourage economic growth in the least developed countries.
Part i) exists in complete working-paper format and will shortly be submitted. I have developed a general equilibrium model, which allows for normative assessment of trade policy as a wealth transfer. The main argument of the original grant application is proven formally (deviations from a nationally optimal trade policy will have only second order effects on nation welfare but generally positive first order effects on foreign welfare and is therefore very generally an efficient way of transferring wealth). The largest deviation from the original grant proposal is in scope where the final model includes political economy considerations, internal equality and multiple countries. This enriches the scope and makes the analysis more broadly applicable. All results have been demonstrated, but the exposition is still incomplete.
Part ii) is still incomplete but a working paper will be submitted in the fall. We have solved some empirical problems and converged on a final empirical strategy and find that the data is broadly consistent with the predictions of our model.
Part ii) partly uses Gravity Equations. In a related paper Pankaj Ghemawat and I show that randomness in productivity draws for firms can accounts for part of the variation in trade flows across countries. The basic argument is as follows: from the perspective of a country entrepreneurial quality is a random draw. A large country, such as the United States, will have many “draws” an as a consequence will have many successful companies spread across several industries. A smaller country, say, Korea, might only have one really successful company that ends up dominating. As a consequence, smaller countries will have more “noisy” distributions of exports across product categories. We have done the formal and empirical analysis and have an early version of a working paper.
Part iii) has progressed substantially more and has in fact given rise to a number of additional project. As the general equilibrium model of part i) is static and takes technology as a given, it can treat the wealth transfers through trade policy but is silent on the ability of trade policy as a means of transferring technology. Such questions are best answered in a model of endogenous technology and we have been developing exactly such a framework. It turns out that the prospective for technological transfers is deeply connected to the progress of technological development in the Western world: Suppose, technological development increasingly substitutes for instead of complementing low-skill labor (Acemoglu and Autor, Handbook of labor economics, 2013) such that the need for low-skill labor in modern production facilities is reduced. When the need for low-skill labor is lower the incentive to move production to low-labor cost countries is lower and the mechanism through which South Korea, China and Japan grew might be weaker in the immediate future than in the second half of the twentieth century. Needless, to say these mechanism will be essential for understanding the ability of trade policy to encourage growth. The intend of the original grant proposal was to extend one of the existing models of endogenous growth to this context. However, we argue that the present theoretical foundations are insufficient to address these issues and instead, along with David Hemous (University of Zurich), I have developed an endogenous growth model of directed technological change which directly addresses the question of when technology might be labor-replacing. This paper is complete and is under review at the American Economic Journal: Macroeconomics.
I have started an additional project (joint with David Hemous and Antoine Dechezlepretre at the London School of Economics) as an empirical investigation of the predictions in the theoretical framework mentioned above. In particular, we use patent codes to classify the universe of patents as either “automation patents” or “non-automation patents” and we ask how the wages facing firms affect these firms' tendency to focus on innovation in automation technology. There is not a yet a working paper, but the preliminary results suggests that 1 per cent higher wages for the low-skill workers employed raises the patenting of automation patents by around 1.1 per cent.
In addition, the progress of technological change is determined by a number of factors specific to the local economy. In a paper with David Hemous, I show that the extend to which firms are engaged in long-term relationships – as a means of overcoming incomplete contract enforcement – determines the direction of technological change. When engaged in strong relationships firms are more likely to focus on relationship specific innovation, whereas economies with less strong business relationships are more prone to larger more general innovations. Using patent data we show that this tendency is borne out by the data. This paper had a Revise and resubmit request from the Journal of Europen Economic Association and has been resubmitted.
From the work on technological change and income inequality sprung a paper on the spill-over effects of income inequality: Suppose technological change or deregulation increases income inequality in a particular occupation, say, financial managers. Then this increase in income inequality will spill over into other occupations such as physicians as the wealthier financial managers compete for the best doctors. We use data from the federal census in the United States to show that this can explain a substantial part of the increase in income inequality. This paper is basically done, but the analysis was carried out on censored Federal Census data. We are presently waiting for the uncensored data. This paper is with David Hemous and Jeffrey Clemens of the University of San Diego.
The expected final result of the project is to assess the potential of trade policy as a means of development aid, and in particular the optimal design of such policies.