Summary of the 4th InsTED / 9th EESP-FGV Workshop

IMG_3897 (2)

We would like to thank the Sao Paulo School of Economics at The Getulio Vargas Foundation (EESP-FGV) for hosting the 4th InsTED Workshop jointly with their 9th Sao Paulo School of Economics Conference.  We are grateful to them, and also to the Swiss Programme for Research on Global Issues for Development, and the “Firms, Markets, and Values” cluster at the University of Exeter Business School, for sponsoring this event.  The workshop took place at EESP-FGV from May 17th-18th 2017.  Special thanks go to our joint-organizers, Emanuel Ornelas and Vladimir Ponczek, as well as Hully Rolemberg for her extremely helpful and patient interactions with everyone.

The program comprised of 18 papers that ranged over four broad topics at the intersection of institutions, trade and economic development.  The first focused on the interactions between international trade and economic development, through the environment, technological change, and institutions.  The second was on the interaction between international trade and labor market institutions, and hence the distributive implications of trade liberalization.  The third concerned analyses of supply chains, their institutional determinants, and how they affect allocative efficiency and hence development.  The fourth was on institutions that shape commercial policy and economic development.  There now follows a summary of all the papers presented at the workshop, organised under these four topic headings.  A bibliography, together with links to papers where available, is provided at the end.  Please note that for brevity the summary mentions presenters’ names but not those of their co-authors.  This information is contained in the bibliography.

Interactions between International Trade and Economic Development

The idea that international trade liberalization goes hand in hand with economic growth and development goes all the way back to Adam Smith’s Wealth of Nations.  But data and research methods to identify the direction of causality between international trade and the various facets of economic development have been developed only relatively recently.  So there is currently an active debate over the directions of causality of these effects and their magnitudes.

The keynote address by Scott Taylor identified an extremely promising new research area focusing on the relationship between international trade and the environment.  He began by taking an overview of the existing body of research in this area, and identified two main concerns: the emergence of pollution havens, and the environmental impact of rapid growth in the developing world driven in part by trade liberalization.  Existing research is based essentially on economy- and industry-level considerations of comparative advantage.  He argued that today’s concerns about the effect of trade liberalization and growth on the environment differ little from those of the past, but the literature has been comparatively slow (relative to other research areas) to adopt a firm-level and plant-level perspective.  Taylor’s presentation then developed tools to facilitate this shift in perspective.  The first tool is a decomposition that allows emissions to be attributed not just at the economy and industry levels, as in past research, but at the more granular firm and plant levels as well.  The second tool is a partial equilibrium model of firm behavior.  The decomposition identifies a set of possible adjustments to trade liberalization that will have environmental implications, while the model allows causal connections to be made between these adjustments.  Finally, Taylor developed a set of new hypotheses to evaluate new environmental predictions derived from models with firm-level heterogeneity. The ‘Pollution Reduction by Rationalization Hypothesis’ links market share reallocations and selection effects in the Melitz model to changes in industry emissions. The ‘Distressed and Dirty Industry Hypothesis’ links changes in abatement and emission intensities to heightened foreign competition brought about by trade liberalization. The ‘Pollution Offshoring Hypothesis’ is a natural analogue to the pollution haven hypothesis of the existing literature, but explicitly links firm level decisions to offshore dirty intermediate inputs to trade liberalization.  Surprisingly, there is as yet no empirical work testing these hypotheses, although a number of existing studies are related.

The keynote address by Amit Khandelwal examined another area of active debate, regarding the relationship between international trade and firm performance in terms of productivity.  There is a long held belief that export-led development strategies improve technical efficiency.  But two difficulties arise with finding evidence of this in the data.  First, what appears to be higher productivity among exporters may in fact be attributable to self-selection by more productive firms into exporting.  Second, information needed to isolate increases in firm-level efficiency arising from trade liberalization is typically not available outside the firm.  In his keynote address, Khandelwal discussed a randomized control trial (RCT) that he and his collaborators had undertaken to address both of these issues.  To address the first, an RCT was undertaken on a group of rug manufacturers in Egypt, whereby the opportunity to export to high-income markets was randomly assigned to some of the firms through a non-governmental organization (NGO) and an Egyptian intermediary.  To address the second, the firms (both those assigned an export opportunity and those who were not) were tested by a skilled quality assessor to measure the quality of the rugs that they produced.  Quality was measured along 11 dimensions, capturing a combination of codifiable specifications and hard-to-codify attributes that depend on the technical skill of the firm.  The results are quite striking.  The opportunity to export raises the profitability of firms by 16-26 percent, arising from the production of higher quality products but at a lower rate.  Adding further nuance to the results, failure to account appropriately for the rise in quality appears to imply that productivity has fallen as a result of the opportunity to export.  In addition, multiple rounds of testing provide evidence of a ‘learning curve’, suggesting an outward movement of firms’ production possibility frontiers (PPFs) rather than a simple shift around the PPF from lower to higher quality rugs.  In conclusion, the application of experimental methods appears to offer compelling evidence that the opportunity to export does indeed increase firm performance.

The program featured two other papers that used experimental methods to assess the effects of trade liberalization.  The paper presented by Tibor Besedes used the eruption of Iceland’s Eyjafjallajökull volcano in 2010 as an unusually clean natural experiment to understand the role that transportation plays in trade and production.  The findings reveal a surprisingly inelastic relationship between this disruption and trade flows.  Air freight volumes from Europe to the US declined by 7 to 10 percent as a result of the week-long disruption, possibly due to the short duration of the airspace closures, but also the flexibility of the airlines in accommodating backlogs.  Michiel Gerritse switched focus to look at the role of trade liberalization in the determination of contracting institutions.  It has been argued that specialization through trade in primary product-based sectors that rely on poor contracting institutions can actually stand in the way of industrial development that requires good contract enforcement.  Gerittse’s paper exploits the closure of the Suez Canal between 1967 to 1975, arising from an unanticipated conflict between Israel and its neighboring Arab nations, as a quasi-natural experiment to assess how trade affects institutional development.  He found that, surprisingly, closure to trade through the Suez gave rise to a fall in institutionally intensive production in affected African nations.  One possible conclusion is that greater openness may give rise to a broader favorable impact on contracting institutions that dominates the corrosive effects on them of specialization in sectors that do not traditionally rely on them.

Reversing the direction of causality, the paper presented by Ana Abelianksky is perhaps the first to consider the effects of 3D printing, regarded as a type of technological innovation, on international trade.  As she explained, the advantage of 3D printing is that it has the potential to significantly reduce the marginal costs of certain types of production.  The drawback is that, especially since it is a relatively new technology, the fixed set-up cost are relatively high.  By seeing 3D printing in this way, she was able to incorporate it naturally into a Helpman-Melitz-Yeaple model, whereby only the most productive firms, and those who face relatively high transport costs, will find it worthwhile to absorb the relatively high fixed costs of 3D printing in order to take advantage of the relatively low marginal costs.  Here paper finds supportive evidence for this in the data.  The intriguing prediction of her model is that as the fixed costs of 3D printing fall, it will eventually displace FDI based on more conventional production line methods, and may ultimately even displace exporting.

Interaction Between International Trade and Labor Market Institutions

When considering the effects of trade liberalization, the interaction with labor markets and the institutions that govern them is of particular interest because these have a critical bearing on the distributional implications.  Joao Paulo Pessoa’s paper captured these interactions by constructing and structurally estimating a dynamic multi-country, multi-sector Ricardian trade model, extended to incorporate both search frictions and labor mobility frictions.  He uses this framework to quantify the effects of China’s integration into the global economy, both in terms of the gains to consumers but also the possible losses to workers through job displacement.  He finds that, while overall welfare gains in Northern countries are positive, in import competing sectors workers bear a costly transition, experiencing lower wages and a rise in unemployment.  The paper presented by Lorenzo Rotunno examined the distributional implications of trade liberalization using a generalized Hecksher-Ohlin (GHO) model.  The generalization arises from imperfect substitutability between home and foreign varieties, and creates inelasticity of demand in labor markets and thus helps explain why relative wages vary with skill supplies in open economies.  In this framework, Rotunno’s paper finds that relative wages in open economies vary with relative skill supplies, as labour economists believe.  But it also finds that the response of wages to variation in skill supplies is smaller in countries with lower barriers to trade, and in very open economies this comes close to the simplest trade economist view that wages are unaffected by endowments.

In his presentation, Sotiris Blanas challenged the perception that foreign firms exploit African workers.  He looked at how variation in country-level institutional factors affect labor market outcomes in Sub-Saharan Africa, differentiating between the quality of jobs offered by foreign-owned and domestic firms.  His paper finds that they foreign-owned firms offer more stable and secure jobs than domestic firms.  The job stability and security advantage of foreign-owned firms is smaller in countries with higher firing costs and governance quality.  But this appears to be because domestic firms are induced to offer more stable and secure jobs, rather than resulting from poorer standards adopted by foreign-owned firms.

Supply Chains: Their Institutional Determinants and Effects on Resource Allocation Efficiency              

An exciting new (or perhaps revived) research topic is how distortions accumulate and are magnified along supply chains, potentially undermining economic development.  Johannes Boehm and Heiwai Tang presented different but complementary extensions of the Hsieh-Klenow methodology to explore this issue.   Boehm’s paper modelled the sophistication of input-output networks across Indian states in terms of variation in the quality of contract enforcement.  The data show that poorer states have slower courts.  To capture distortions, he structurally estimates ‘wedge parameters’ from firms’ expenditure shares on region-specific intermediate inputs.  He finds that the size of the identified wedges on relationship-specific intermediate inputs is strongly correlated with the length of backlogs in regional courts.  His estimates for the wedges imply that the costs of misallocation along supply chains are economically significant.  Moreover, some of this misallocation is attributable to the slow enforcement of contracts in the courts, implying that reforms aimed at reducing court congestion could have large effects on aggregate productivity.  The paper presented by Tang studies firms’ decisions to source inputs, but focuses on the role of global versus domestic sourcing decisions in the determination of total factor productivity (TFP).  Solving the general equilibrium model with industry linkages reveals that an economy’s aggregate TFP losses due to distortions should be equal to the geometric mean of sector-level TFP losses, implying that weights are equal to the sectors’ Domar weights (i.e., the ratio of gross output to total manufacturing value added).  The fact that the sum of the Domar weights across sectors is always larger than one implies that the aggregate TFP loss due to resource misallocation may have been underestimated in the literature.  Using this approach, he finds that China’s TFP losses are smaller than those for India, which seems significant for their respective development paths.

There is a growing consensus that, to understand the implications of trade policy, one must look not so much at its effects on final goods but on traded intermediates.  Emanuel Ornelas showed, in his paper, that preferential trade agreements (PTAs) can help to overcome weak contract enforcement over inputs, and hence tackle inefficiencies that arise along supply chains.  All else equal, PTA trading partners share a higher surplus on every unit traded, relative to what they could obtain by dealing with alternative producers in non-member countries.  This propels firms to trade more, which in turn induces them to increase their relationship-specific investments in inputs.  Since the investment yields greater value to every unit traded, this relationship-strengthening effect is stronger, the more units the firms initially trade.  Since without the PTA there is underinvestment due to a hold-up problem, the PTA-induced investment will generally improve efficiency.  This beneficial effect is more likely to overcome any negative effect of tariff discrimination the higher are trade volumes initially.

Most economists have had their attention drawn to Special Economic Zones (SEZs) through their extensive use by China in recent years.  But Matthew Grant made the compelling case that SEZs are much more prevalent around the world than most would have realized, and provided a framework for their analysis.  Their key feature is that they can be used to discriminate across importers of intermediate inputs.  Grant provides a theoretical framework in which tariff discrimination across importers is optimal policy for a government motivated by both political and welfare considerations. He shows that optimal policy follows a simple two-tiered tariff rule, in which some importers are charged the prevailing tariff, and other firms are charged a reduced tariff. This policy is implemented in practice through selective permission to produce in SEZs. Using a novel data set that he constructed from public records covering the universe of active SEZs in the United States, he shows that the model’s predictions about the size and industrial composition of SEZs are consistent with the way they are implemented in practice.

Rick Bond’s presentation about the ‘destination based cash flow tax’ further reinforced the idea that the effects of policy on trade should increasingly be understood in terms of its effects on intermediates.  As Bond’s presentation explained, the implementation of this policy has been referred to as a ‘border tax adjustment’ because it exempts export sales from the tax but does not allow firms to deduct purchases of imported intermediates from the taxable cash flow.  The aim of this tax, which would replace the current US corporate tax system, is to remove the incentive of multinational firms to locate part of their operations in low-tax jurisdictions.  Proponents of the tax have argued that the distributional implications are neutral.  But Bond showed that under plausible assumptions of a specific-factors model the tax would effectively raise the return to import-competing goods relative to exportables, resulting in a flow of resources out of the exportable sector.

Institutions That Shape Commercial Policy and Economic Development

Considerable strides have been taken over the last twenty years to provide an ‘optimal trade policy rationale’ for the World Trade Organization, as an institution whose members have an incentive to engage in mutual trade liberalization in the face of a unilateral incentive to protect.  Over a similar timeframe, the gravity model has become the mainstay of empirical work on international trade, but has made very little contact with the literature on optimal trade policy.  The paper presented by Mostafa Beshkar takes a decisive step to build a bridge between these literatures.  The model of the paper is based on an Eaton-Kortum-Armington framework, making it possible to study trade policy across sectors that have very different trade elasticities and degrees of product differentiation.  A striking result in this framework is that import and export policies are complements rather than being substitutes as per the Lerner-Symmetry theorem.  This provides a new rationale for ‘banning’ export policies at the WTO because this provides an immediate impetus to lower import tariffs.

The rising tide of nationalism around the world has highlighted the inherent uncertainty in the ratification of trade agreements in domestic legislatures.  The paper presented by Ben Zissimos develops a framework that puts the uncertainty of ratification at the center of the stage.  To create ratification uncertainty, the framework features a contest that takes place in each country between interest groups who are ‘for’ and ‘against’ the agreement respectively.  This framework extends the traditional contest framework wherein there is only one decision-maker, in this context a national government, to where there can be more than one.  A key prediction of the framework is that lobbying drives trade liberalization, while it is governments’ protectionist concerns (not those of lobbies) that hold liberalization back.  This reverses the logic that prevails in the literature but is consistent with recent econometric findings.

On a related theme, Kristy Buzard asked whether there could be a motive for the gradual reduction of trade policy through trade agreements that relied purely on political-economy motives.  Previous papers in the gradualism literature have tended to focus on underlying features of the economy rather than politics.  These include stickiness in the movement of factors between sectors, or stickiness in the adjustment of trade policy introduced by WTO rules.  Buzard’s motivation for gradualism is a political economy shock, leading to the ousting from office of a key politician on whom protectionist policy relies.  This gives rise to a ratcheting effect whereby resources move out of the import-competing sector, yielding fewer resources to support politicians and, in turn, further losses of office and further reductions in protection.

While economic globalization, supported by the WTO, is leading to a liberal markets around the world, Bernardo Guimaraes argued that ‘politics seems to be immune from this trend’.  The ‘liberalization’ of political institutions is more limited, and while the rule of law can be taken for granted in many places, authoritarianism is still prevalent despite its negative consequences.  The paper that he presented develops a theory of political specialization to understand how an increasingly interconnected world can nonetheless sustain diametrically opposed systems of government.  According to the theory, some countries will uphold the rule of law with a commitment to property rights, while others will consciously choose not to do so.  This political specialization relies on an interplay between two key factors: diminishing marginal benefits to good government at the world level but not at the country level; and diminishing marginal costs of good government at the country level.  The theory implies that political specialization is to be expected even if all countries are ex ante identical, which means good governance everywhere is a remote prospect.

The remarkable thing about the keynote address by Gianmarco Ottaviano, the final presentation, was the way that it spanned so many of the topics that had been discussed throughout the workshop.  The paper that he presented asked how multilateral trade policy should be designed in a world with two key features.  First, countries differ in terms of market access and technology.  And second, firms with market power differ in terms of productivity.  The framework he developed to address this question extends the Melitz-Ottaviano framework to one of multilateral trade, in which variable markups that increase in firm size are a key source of misallocation across firms and countries.  The framework makes it possible to answer a number of key questions.  For example, how should multilateral trade policy be designed in a world in which firms with market power differ in terms of productivity?  Should worse performing (national) firms be protected from better performing (foreign) rivals? Should national product diversity be shielded against the potentially disruptive effects of cheaper imported goods?  In the canonical models of this literature, based on CES demands, constant marginal costs, and a Pareto distribution for firm productivities, free trade is efficient and multilateral trade policy ‘should’ reflect this.  The main purpose of the paper is to show that this efficient outcome ceases to hold when the CES assumption is removed, giving rise to new implications for multilateral trade policy aimed at maximizing the joint welfare of all trade partners.  To do this, the paper focuses on demands that satisfy Marshall’s Second Law of Demand, according to which demand becomes more inelastic with consumption.  The outcomes under this comparatively modest modification are strikingly different.  First, from a welfare point of view, too large a range of products is sold to larger markets, while too small a range is sold to smaller markets.  Second, conditional on range, relatively too many high cost products are sold to any country. This inefficiency is, however, more severe for smaller countries. Third, conditional on range and selection, the quantities of high cost products sold to any country are too large and those of low cost products are too small.  Also, this inefficiency is more severe for smaller countries. As a result, the free market provides an inefficiently high degree of welfare inequality between large and small countries. There is, therefore, room for welfare improving multilateral policy intervention that: increases sales of low cost firms to all countries but especially to smaller ones; decreases sales of high cost firms to all countries but especially to smaller ones; reduces firm entry in all countries but especially in smaller ones.  This work has important implications for the optimal trade policy literature in terms of introducing considerations of firm heterogeneity that have not been discussed previously.  It also has implications for work based on the Hseih-Klenow methodology by providing a more nuanced perspective on the estimation of wedges in a world where demand is non-CES, and therefore free trade does not necessarily increase welfare.

 Bibliography of Papers Presented with Links Where Available (Presenters’ Names Shown in Bold)

Abeliansky, A., I. Martínez-Zarzoso, and K. Prettner, “How Does 3D Printing Affect Globalization?

Atkin, D., A. Khandelwal, and A. Osman, “Exporting and Firm Performance: Evidence from a Randomized Experiment.

Besedes, T., and A. Murshid, “Experimenting with Ash: The Trade-Effects of Airspace Closures in the Aftermath of Eyjafjallajökull.

Beshkar, M., and A. Lashkaripour, “Interdependence of Trade Policies in General Equilibrium.

Bickwit, G., E. Ornelas, and J. Turner, “Preferential Trade Agreements and Global Sourcing.

Blanas, S., A. Seric, and C. Viegelahn, “Jobs, FDI and Institutions in Sub-Saharan Africa: Evidence from Firm-Level Data.

Boehm, J., and E. Oberfield “Misallocation in the Market for Inputs.”

Bond, E., and R. Driskill, “DBCFT, Border Adjustments, and Trade.”

Buzard, K., “Explaining Gradualism in Trade Liberalization: A Political Economy Approach.”

Cherniwchan, J., B.R. Copeland, and S. Taylor, “Trade and the Environment: New Methods, Measurements, and Results.”  CESifo working paper

Cole, M., J. Lake, and B. ZissimosContesting an International Trade Agreement.

Gerritse, M., “Does Trade Cause Unfortunate Specialization in Developing Economies? Evidence from Countries South of the Suez Canal.”

Grant, M., “Why Special Economic Zones? Using Trade Policy to Discriminate Across Importers.

Guimarães, B., and K. Sheedy “Political Specialization.

Krishna, P., and H. Tang, “Production Networks and Misallocation.

Nocco, A., G. Ottaviano, and M. Salto “Geography, Competition, and Optimal Multilateral Trade Policy.”

Pessoa, J.P., “International Competition and Labor Market Adjustment

Rotunno, L., and A. Wood, “Wage Inequality and Skill Supplies in a Globalized World