 |
| Stefan Beck |
 |
| Christoph Scherrer |
The US government and the European Commission are negotiating a Transatlantic Trade and Investment Partnership (TTIP). The trading partners push for the elimination of the tariffs, less regulation, and more rights for investors. The governments try to justify the TTIP by pointing to significant welfare gains from additional exports, higher growth, and increases in efficiency, income and employment. They draw on various economic studies which, are based on complex modelling; concluding that a TTIP would increase prosperity in both the US and the European Union.
Trade impact studies usually make use of Computable General Equilibrium models. These models have been criticised in recent years. Among the critics is a team from the Munich ifo Institute which produced two impact studies in 2013, one for the German Federal Ministry of Economics (ifo-BMWi) and the other for the Bertelsmann Foundation (ifo-Bertelsmann). We concentrate our assessment on the impact studies drawing from the ifo studies.
Trade impact studies usually make use of Computable General Equilibrium models. These models have been criticised in recent years. Among the critics is a team from the Munich ifo Institute which produced two impact studies in 2013, one for the German Federal Ministry of Economics (ifo-BMWi) and the other for the Bertelsmann Foundation (ifo-Bertelsmann). We concentrate our assessment on the impact studies drawing from the ifo studies.
The TTIP Scenarios of the ifo-Studies
The ifo studies distinguish between various scenarios of the outcome of the TTIP negotiations. Two scenarios are highlighted: a “tariff scenario”, which assumes the complete dismantling of the remaining tariffs, and a “NTB scenario” (ifo-BMWi) or “comprehensive agreement” (ifo-Bertelsmann) which assumes complete elimination and a comprehensive reduction of non-tariff barriers. The ifo-BMWi study calculated for the “tariff scenario” a total increase in real wages in Germany of 0.13 %, the ifo-Bertelsmann-study is more optimistic (0.54 %) and expects 45,000 additional jobs for Germany. In its “NTB scenario” the ifo-BMWi-study calculated 25,220 additional jobs, while the ifo-Bertelsmann-study leads to an even higher number of jobs: 181,000 (s. Table1). All of these gains are higher than those reported by the studies using Computable General Equilibrium models.
Critique of the economic benefit calculations
The growth/welfare effects in both ifo studies crucially depend on the assumption that TTIP will increase foreign trade between participating countries by about 80%. According to the authors, free trade agreements which are already in force have increased foreign trade between participating countries by this amount (on average) in the past. Since this is a very strong assumption, it should be discussed in more detail.
Econometric estimates promise a better reflection of reality, but the question is whether that promise would hold into the future. A first point to be clarified is how far the preferential trade agreements implemented up to now are comparable. The study compares all agreements registered with the WTO before 2005. It is targeted on long-term effects. A long time frame is specified three times over in the study – but the length is different each time: 5-8 quarters (p. 14 fn13), 10-20 years (p. 69) and 12 years (p. 111). The greater the timespan, the more likely it becomes that the impacts measured will be influenced by other factors. The regression analysis in the study was made for a particular point in time (cross-sectional analysis) and was not checked for the whole duration of the agreements, it remains uncertain whether the agreements analysed are comparable. To date, no trade agreement exists that is comparable to the TTIP in terms of the size of the economic area.
A more fundamental question, given the negative experiences during the financial crisis with forecasting models based on data from the past, is whether past data in general can just be casually extended into the future. Technical progress makes it difficult to see into the future. How can the part played by a free trade agreement in increasing productivity be precisely measured against other factors? Even more basically, can the impact of one single factor ever be separated out in complex systems such as economies? That would imply that the impact of the factor under investigation was not dependent on its specific interaction with each other factor. For instance, an agreement on cross-border education would have had a different impact in the pre-internet era.
Measuring non-tariff trade barriers (NTB)
The ifo-BMWi study concedes that the determination of the effects of non-tariff trade barriers (NTB) by analogy to past free trade agreements provides results “which are informative, but only partially reliable” (p. 42; translated ChS). It takes the step of complementing its econometric analyses with the results of a survey of German business associations concerning trade costs in the US. However, only 16 associations replied, mainly from the associations that had already advocated a TTIP. Nonetheless, some of the answers are interesting. For instance, US anti-terrorism laws were cited as barriers. It is unclear how far this legislation will be altered to accommodate the TTIP. Also particularly mentioned were sectoral barriers. In the case of the finance industry, these are formed by the complex legislation adopted in response to the 2007 financial crisis, the Dodd-Frank Act, and for the food industry. The study does not raise the question of whether such provisions are reasonable and should therefore not be swept aside by the TTIP.
The study sees the survey as justifying its decision to base the calculations on a goods-related economy and to set the currency aspects aside, as “the exchange rate issue [was] not regarded as very important by the interviewees […] except in the textile sector and vehicle manufacturing” (p. 55). Elsewhere, though, the study singles out these two sectors, textiles and vehicle manufacture, as standing to profit most from the TTIP.
The study assumes in its econometric analysis of non-tariff barriers between the EU and the US that there are no trade barriers within the US, within the North American Free Trade Agreement (NAFTA) zone and especially within the EU. However, even inside the US trade barriers do exist. By not taking account of internal barriers, the level of supposed external barriers is raised. The study calculates that the “imputed trade costs” for Germany’s exports to the US are 53% higher than in Germany. US exports to Germany, meanwhile, are 155% more expensive than in the US. The study explains away this difference “mainly through the fact that ... Germany achieved high bilateral surpluses in 2007” (p. 89). As Germany’s surpluses subsequently showed even stronger growth, it follows that barriers to trade into Germany must have also increased. If the level of trade barriers is measured in this way the implication is that trade barriers vary from year to year, in line with the balance of foreign trade, which is not plausible.
Employment impacts
The ifo-BMWi study attempts to introduce more realistic assumptions into the models for calculating the effects of trade agreements. When investigating the employment impacts, it takes account of the initial employment levels, job-seeking rates and the respective labour market institutions.
As trade-induced employment growth is fuelled by falling unemployment and job losses in uncompetitive enterprises, taking account of the level of unemployment produces higher employment impacts for free trade. As the study puts it, “it is even the case that lower unemployment figures in the basic balance would lead to even smaller potential improvements through the free trade initiative” (p. 99, footnote 51). At the same time it points out “that the results of this study do not depend on the level of unemployment in Germany” (ibid.).
From the study, it is not possible to discover how it treats frictional unemployment. From empirical studies of the job search processes used by workers who have been displaced by international competition or technological progress, it can be seen that in both the US and Germany, a significant proportion of unemployed people experience long search times and usually have to accept lower pay from the new employer.
The estimated employment gains look big in absolute numbers (see Table 1) but pale in relation to the overall labour pool. Under the “realistic” scenario, 25,000 new jobs would be created in total in Germany. That is just about half a per mille of the 41.8 million people gainfully employed in 2012.
Moreover, they assert that trade-induced productivity increases would make it possible to increase wages: in Germany, an average of about €50 in additional gross monthly earnings. These wage increases would be “very strongly driven by the price index” (p. 104), which in turn would be powered by trade-driven productivity developments that “lead to a reduction in average prices to domestic consumers” (p. 99). A small part of rising wages are supposedly due to shifts of employment to more productive firms.
Conclusion
Interestingly, the word “forecast” never crops up in the ifo-studies. As is usual for scientific forecasts, the study works with scenarios, but these relate only to the presumed extent of liberalisation due to the TTIP. For each stage of liberalisation, just one impact scenario is presented. Those who commissioned these studies had already come down in favour of the TTIP, so it is scarcely surprising that they should ignore the possible caveats surrounding a scientific study. Sadly, these ifo-studies have made it easy to take that attitude. The forecasts of economic impact studies should always deserve close scrutiny. In most cases their results will be driven by their (biased) assumptions.
Table 1: Impacts on Germany, ifo-BMWi und ifo-Bertelsmann, scenarios
 |
| Note: figures from the ifo-Bertelsmann study are in italics. Sources: Felbermayr et al. 2013a: 100 and 2013b: 36-41. |
Ifo studies apply their own methods in an attempt to avoid three of the weaknesses they identified in the equilibrium models typically used. They attempt to base the parameters of the models on economic estimates that reflect reality better than theoretical deductions. These econometric estimates relate to the trade impacts of comparable free trade agreements. The study does not presuppose full employment, nor does it assume perfect competition.
The growth/welfare effects in both ifo studies crucially depend on the assumption that TTIP will increase foreign trade between participating countries by about 80%. According to the authors, free trade agreements which are already in force have increased foreign trade between participating countries by this amount (on average) in the past. Since this is a very strong assumption, it should be discussed in more detail.
Econometric estimates promise a better reflection of reality, but the question is whether that promise would hold into the future. A first point to be clarified is how far the preferential trade agreements implemented up to now are comparable. The study compares all agreements registered with the WTO before 2005. It is targeted on long-term effects. A long time frame is specified three times over in the study – but the length is different each time: 5-8 quarters (p. 14 fn13), 10-20 years (p. 69) and 12 years (p. 111). The greater the timespan, the more likely it becomes that the impacts measured will be influenced by other factors. The regression analysis in the study was made for a particular point in time (cross-sectional analysis) and was not checked for the whole duration of the agreements, it remains uncertain whether the agreements analysed are comparable. To date, no trade agreement exists that is comparable to the TTIP in terms of the size of the economic area.
A more fundamental question, given the negative experiences during the financial crisis with forecasting models based on data from the past, is whether past data in general can just be casually extended into the future. Technical progress makes it difficult to see into the future. How can the part played by a free trade agreement in increasing productivity be precisely measured against other factors? Even more basically, can the impact of one single factor ever be separated out in complex systems such as economies? That would imply that the impact of the factor under investigation was not dependent on its specific interaction with each other factor. For instance, an agreement on cross-border education would have had a different impact in the pre-internet era.
Measuring non-tariff trade barriers (NTB)
The ifo-BMWi study concedes that the determination of the effects of non-tariff trade barriers (NTB) by analogy to past free trade agreements provides results “which are informative, but only partially reliable” (p. 42; translated ChS). It takes the step of complementing its econometric analyses with the results of a survey of German business associations concerning trade costs in the US. However, only 16 associations replied, mainly from the associations that had already advocated a TTIP. Nonetheless, some of the answers are interesting. For instance, US anti-terrorism laws were cited as barriers. It is unclear how far this legislation will be altered to accommodate the TTIP. Also particularly mentioned were sectoral barriers. In the case of the finance industry, these are formed by the complex legislation adopted in response to the 2007 financial crisis, the Dodd-Frank Act, and for the food industry. The study does not raise the question of whether such provisions are reasonable and should therefore not be swept aside by the TTIP.
The study sees the survey as justifying its decision to base the calculations on a goods-related economy and to set the currency aspects aside, as “the exchange rate issue [was] not regarded as very important by the interviewees […] except in the textile sector and vehicle manufacturing” (p. 55). Elsewhere, though, the study singles out these two sectors, textiles and vehicle manufacture, as standing to profit most from the TTIP.
The study assumes in its econometric analysis of non-tariff barriers between the EU and the US that there are no trade barriers within the US, within the North American Free Trade Agreement (NAFTA) zone and especially within the EU. However, even inside the US trade barriers do exist. By not taking account of internal barriers, the level of supposed external barriers is raised. The study calculates that the “imputed trade costs” for Germany’s exports to the US are 53% higher than in Germany. US exports to Germany, meanwhile, are 155% more expensive than in the US. The study explains away this difference “mainly through the fact that ... Germany achieved high bilateral surpluses in 2007” (p. 89). As Germany’s surpluses subsequently showed even stronger growth, it follows that barriers to trade into Germany must have also increased. If the level of trade barriers is measured in this way the implication is that trade barriers vary from year to year, in line with the balance of foreign trade, which is not plausible.
Employment impacts
The ifo-BMWi study attempts to introduce more realistic assumptions into the models for calculating the effects of trade agreements. When investigating the employment impacts, it takes account of the initial employment levels, job-seeking rates and the respective labour market institutions.
As trade-induced employment growth is fuelled by falling unemployment and job losses in uncompetitive enterprises, taking account of the level of unemployment produces higher employment impacts for free trade. As the study puts it, “it is even the case that lower unemployment figures in the basic balance would lead to even smaller potential improvements through the free trade initiative” (p. 99, footnote 51). At the same time it points out “that the results of this study do not depend on the level of unemployment in Germany” (ibid.).
From the study, it is not possible to discover how it treats frictional unemployment. From empirical studies of the job search processes used by workers who have been displaced by international competition or technological progress, it can be seen that in both the US and Germany, a significant proportion of unemployed people experience long search times and usually have to accept lower pay from the new employer.
The estimated employment gains look big in absolute numbers (see Table 1) but pale in relation to the overall labour pool. Under the “realistic” scenario, 25,000 new jobs would be created in total in Germany. That is just about half a per mille of the 41.8 million people gainfully employed in 2012.
Moreover, they assert that trade-induced productivity increases would make it possible to increase wages: in Germany, an average of about €50 in additional gross monthly earnings. These wage increases would be “very strongly driven by the price index” (p. 104), which in turn would be powered by trade-driven productivity developments that “lead to a reduction in average prices to domestic consumers” (p. 99). A small part of rising wages are supposedly due to shifts of employment to more productive firms.
Conclusion
Interestingly, the word “forecast” never crops up in the ifo-studies. As is usual for scientific forecasts, the study works with scenarios, but these relate only to the presumed extent of liberalisation due to the TTIP. For each stage of liberalisation, just one impact scenario is presented. Those who commissioned these studies had already come down in favour of the TTIP, so it is scarcely surprising that they should ignore the possible caveats surrounding a scientific study. Sadly, these ifo-studies have made it easy to take that attitude. The forecasts of economic impact studies should always deserve close scrutiny. In most cases their results will be driven by their (biased) assumptions.
Stefan Beck holds a PhD in political science. He has recently completed a comprehensive study on TTIP.
Christoph Scherrer is Professor for Globalization and Politics at University of Kassel, Germany. He is also executive director of the International Center for Development and Decent Work and a member of the steering committee of the Global Labour University.
Christoph Scherrer is Professor for Globalization and Politics at University of Kassel, Germany. He is also executive director of the International Center for Development and Decent Work and a member of the steering committee of the Global Labour University.
References
CEPR (2013): Reducing Transatlantic Barriers to Trade and Investment. An Economic Assessment, Centre for Economic Policy Research, London,
ETUC position on the Transatlantic Trade and Investment Partnership
Felbermayr, G./Larch, M./Flach, L./Yalcin, E./Benz, S. (2013a): Dimensionen und Auswirkungen eines Freihandelsabkommens zwischen der EU und den USA, Studie im Auftrag des BMWi, IFO Institut, München.
0 Comments:
Post a Comment