Study: Who Will Benefit in Germany from a TTIP?

States, Branches of Industry and
Education Levels
Who will Benefit in Germany from a Transatlantic Trade
and Investment Partnership (TTIP)?
Final Report, Part 2: Microeconomic Effects in Germany

States, Branches of industry, and
Education Levels
Who Will Benefit in Germany from a Transatlantic Trade
and Investment Partnership (TTIP)?
Final Report, Part 2: Microeconomic Effects in Germany
Prof. Gabriel Felbermayr, Ph. D.
Sybille Lehwald
Dr. Ulrich Schoof
Mirko Ronge

5
Table of Contents
Table of Contents
1. Introduction 6
2. Comments on the investigational method 7
2.1 Sector trade effects 7
2.2 Trade effects for occupational groups, education levels and states 8
2.3 Effects on real wages and wage disparities 8
3. Data and trends 10
3.1 Trade data 10
3.2 Input-Output Dataset 11
3.3 Regional Data 12
3.4 Company Data 12
3.5 Wage data 12
4. How are the trade effects distributed across industries? 15
4.1 Trade creation and NTB quantification 15
4.2 Indirect effects on the service sectors 17
4.3 Value creation and employment effects 20
5. Impact of education and occupational groups and impact of regions 22
5.1 Descriptive analysis of the education groups 22
5.2 Impact of education groups 25
5.3 Descriptive analysis of the occupational groups 25
5.4 Impact on occupational groups 26
5.5 Descriptive analysis of the regions 26
5.6 Impact on regions 28
5.7 Value creation and employment effects in the regions 29
5.8 Employment and education in the regions 30
6. Effects of a TTIP on income and the income risk 32
6.1 Effects on real income 32
6.2 Effects on income risk 34
7. Summary 36
A. Annex 37
A.1 Gravitation models, estimation methods and results 37
Literature 48

6
1. Introduction
1. Introduction
In the first part of this study we examined the macroeconomic effects of a transatlantic trade and
investment partnership (TTIP) between the European Union and the USA1. The main focus was on
changes in trade structure, real income and employment provoked by the TTIP. Using a general
equilibrium model, aggregated effects were examined for more than 120 countries. Adaptation
of the aggregated price index in all these countries and the feedback effects on gross domestic
products were considered in doing so, as was the full matrix of trade effects (even those between
countries only indirectly affected) and thus all worldwide trade diversion effects. However, the
broad geographic scale and the focus on macro-economic results made it impossible to draw more
precise conclusions within individual countries. The second part of the study is intended to close
that gap for Germany.
This study segment is devoted to the microeconomic effects of a transatlantic trade and investment
partnership. It consists of zooming in on Germany, where we examine the disaggregated effects
of an agreement on Germany’s sectors and regions. This framework makes it possible to clarify
which sectors and regions would be more impacted by a potential trade agreement than others.
Furthermore, we can analyze the effects of a TTIP on different education levels and occupational
categories. Unlike Part 1 of our study, we do not apply a general equilibrium model. Instead we
base our analysis on a partial analytical approach of a gravitation equation. This means that the
effects of the general equilibrium, such as trade diversion effects, are left out. This circumstance
is important when interpreting the results and has obvious implications for any comparison with
the results of the macro study.
Our analysis makes it possible to identify differences for Germany in how the impacts of a TTIP
are felt in particular industries, occupations and education categories or regions. While our study
makes use of quantitative methods, its partial analytic nature study strongly suggests that the
results be interpreted mainly qualitatively. For a discussion of the effects on the whole economy,
we refer the reader to the macro study cited above.
This micro study is organized as follows: In section 2 we explain our methodological approach and
then, in section 3, we briefly present the data we used. In section 4, we identify the sector trade
effects and discuss which industries we anticipate will experience greater economic vitality than
others. Moreover, in this section we use a new approach to quantify the extent of non-tariff barriers
(NTBs) at the industry level. Beyond that, we calculate the value creation and employment effects
at the sector level. In section 5, we look at both the effects of a TTIP on Germany’s job market as
well as on Germany’s regions. We explain which occupation and education categories and which
regions would be most impacted by such an agreement. In section 6 we look at the effects of a TTIP
on real income and the income risk. At the end of the study, we summarize our results.
1 See “The Transatlantic Trade and Investment Partnership (TTIP). Who Benefits from a Transatlantic Free Trade Agreement?
Part 1: Macroeconomic Effects.“

7
2. Comments on the investigational method
Our analysis is divided into three major steps. First we estimate the effect of a transatlantic
agreement on trade between the USA and Germany in 16 different economic sectors on the
basis of gravitational equations. For these economic areas, we have robust data and can readily
estimate the projected induced trade effects.2 This gives us indicators that reflect the impact on
specific sectors of a TTIP. In the second step, we apply these sector impact levels to 88 different
occupational groups, 3 education levels and 16 regions within Germany’s federal states. In a third
step, we use these results to estimate the effects of a TTIP on real wages in individual occupational
groups and education levels. We also then calculate the effect of the TTIP on wage distribution.
2.1 Sector trade effects
The starting point for the first step is to estimate the expected trade effects at the sector level.
As we did in the first part of the study, we assume that a transatlantic agreement will result in
trade-creation effects that are similar to those for which data already exists. The main distinction
from the first part of this study is the sector disaggregation: We estimate the extent to which the
free trade agreement has led to trade creation within the affected country pairs for 16 different
industries and then use the result as the most credible estimator for the effects of a transatlantic
agreement.3
This approach has the major advantage of allowing a simple quantification of the potential effects
of the agreement on non-tariff barriers. In that way, besides the elimination of import duties, which
we can take for granted, we can include all important categories of trade costs whose decline would
result in stimulating trade between the USA and the EU. Specifically, our approach considers all
costs that limit international trade between two countries but do not fall in the category of import
duties. Non-tariff barriers are regulatory measures with protectionist effects that disadvantage
foreign suppliers compared to domestic ones. They can be politically induced or result from
geographic and historical circumstances.
This economic analysis based on the gravitation model provides us with benchmarks for the
increase in trade between Germany and the USA expected from eliminating customs barriers and
non-tariff trade costs. In terms of methodology, we use a completely saturated fixed effect model on
panel data for annual industry data from 1998 to 2007 and thereby exclude the recent crisis years.4
We are especially interested in the average effects of free trade agreements on trade,5 because
2 The 16 sectors consist of 14 in the manufacturing sector plus agriculture and mining. They are described below as manufacturing
sectors. For the service sector, we calculate indirect effects through inter- and intra-sector links.
3 Another difference consists in the underlying model frameworks. Unlike Part 1, we do not use a general equilibrium model but
a partial analytic model.
4 Such a saturated fixed effect model has been used by Felbermayr and Yalcin (2013), for example. See the references to the
literature cited there. In Annex A.1 to this study, we explain the estimation method in greater detail.
5 There is a special challenge to empirical modelling in trying to provide coverage that is as complete as possible of all potential
trade flow determinants. Only when that is assured can the effects of a free trade ageement be isolated and treated as causal.

8
they can be used to draw conclusions about changes in trade costs at the sector level. Moreover,
because the average customs duties applied are known, the significance of non-tariff barriers
(more exactly: the expected extent of their reduction) can be quantified.
Based on quantified trade potentials in the manufacturing industries and effects that can be
interpolated from them for the service sector, we show where the largest value creation effects can
be expected and in which industries employment will be most affected by a TTIP.
2.2 Trade effects for occupational groups, education levels and
states
We begin by estimating the expected trade effects of a TTIP at the industry level. For that we
use official trade statistics, as described in greater length below. We translate the trade shocks
identified in this way, using data from the Institute for Employment Research (IAB) in Nuremberg,
into shocks for individual occupational groups and education levels.
The approach is as follows: We know from the IAB dataset how various occupational groups
are distributed among the individual segments of the economy, or what share of employment
is held within specific industrial sectors by members of various occupational groups. The same
applies to the different levels of education (university degrees, high school diplomas and/or
vocational training or less). From the interaction of the trade shocks identified in step one with
the employment distributions described, we can convert them into shocks specific to occupational
groups and education levels. We have selected an appropriate ap-proach to quantify the shock
at the state level using regional foreign trade statistics from the Federal Statistical Office, which
provides information about export and trade activities in the individual sectors throughout every
state in Germany. This enables us to draw conclusions about regional industry effects.
2.3 Effects on real wages and wage disparities
In the final step, we use the shocks for occupational groups and education levels described
above in Mincer wage equations. Such equations model workers’ wages as a function of their
characteristics. To conduct our analysis, we have expanded the classic model to include employer
characteristics. What interests us in particular is the extent to which the establishment in which a
worker works is affected by international trade (exports, imports). The literature typically reports
that companies with a more pronounced international presence pay higher wages than those that
are less internationally oriented or are active only on the domestic market.6
6 Felbermayr, Hauptmann and Schmerer (2013) discuss methodological aspects of the estimates of such equations and the
classification of the results in the literature on trade and labor market theory.

9
We can now use the shocks calculated by the estimated Mincer equations in the first and second
steps of our analysis to project average changes in real wages resulting from a TTIP. This analysis
can be executed on samples so that the real wage effects can be identified in individual segments
of the German labor market (occupational groups and education levels).
Similarly, using the means described above, it is possible to forecast changes in wage disparities
among individual segments of the German labor market. To do that, the individual data must be
aggregated, however. This is done by calculating a wage disparity benchmark for the labor market
segment under consideration. As usual in the literature, we use the standard deviation of the
logarithm of wages. This benchmark does not depend on scaling the wage variables and is related
monotonically to the well-known Gini Index of income inequality. Accordingly, a higher standard
deviation indicates a higher degree of inequality.

10
3. Data and trends
3. Data and trends
3.1 Trade data
Our analysis is based on bilateral trade data at the industry level. We use the BACI dataset
developed by CEPII with UN COMTRADE data and thus including trade information on all UN
countries.7
Unfortunately, there are narrow limits to disaggregation by sector. This is due on the
one hand to the fact that we must ensure that our sectors are recognized in the system used for
the input-output tables, and on the other, that our sector classification is compatible with the IAB
datasets. Our sector classification is based on the standard classification of economic activities in
the European Union (NACE Rev. 1.1) defined to two places. For our analysis of the trade effects
induced by a TTIP, we examine 16 manufacturing sectors. These account for about 87 percent of
German foreign trade, with only 13 percent included in the service sector.8 The sectors we examine
thus cover the greater majority of all of Germany’s direct trade relationships. This fact, combined
with the poor data situation for trade in services in general, justifies concentration on the 16
sectors in our analysis of the trade effects induced by a TTIP.
Figure 2 shows the average annual rate of change of trade between Germany and, respectively, the
EU (defined as the EU 27), the USA, the BRICS countries (Brazil, Russia, India, China and South
Africa) and the whole world for the sectors we examined. It becomes clear that in every sector
except petroleum, there was less change in German trade with the USA than with the BRICS
countries or with the whole world. This may be due to the fact that the level of trade with the USA
is already substantially higher than with the emerging countries, where there is need to catch up
with the USA. However, trade with the USA has increased in 12 of the 16 sectors less than it has
with the EU, due to the number of new countries joining the customs union and the extension of
the domestic market program. Nevertheless, this finding clearly shows that there is a potential
for more trade between the EU and the USA from eliminating customs and regulatory barriers to
market entry.
7 For more information, see: http://www.cepii.fr/CEPII/en/bdd_modele/presentation.asp?id=1.
8 The basis for these calculations is data from the World Input Output Tables from 2007. This was the last year before the sharp
decline in world trade as a result of the 2008 and 2009 financial market crisis.

11
3. Data and trends
3.2 Input-Output Dataset
The 16 manufacturing sectors we studied account for more than 80 percent of German foreign
trade, but account for only some 20 percent of total value creation and about 20 percent of all
employees.9 In order to be able to estimate the effects of a transatlantic agreement for the whole
economy, we additionally calculate the indirectly induced effects for the service sector. To do this,
we use the inter- and intra-sector interconnections of the input-output analysis, based on the World
Input-Output Tables (WIOD).10
In the text that follows, we therefore distinguish between the direct effects of a TTIP, which result
in changes in the trade volume in the 16 manufacturing sectors, and the indirect effects of a TTIP,
induced by interconnecting relationships.
9 These calculations are based on the data in the World Input-Output Tables form 2007. See also footnote 19. According to the most
recent data from the German Federal Statistics Office from 2012, manufacturing generates 31.6 percent of all value creation and
26.6 percent of all employees.
10 For more information, see http://www.wiod.org/index.htm.
Source: Calculations by the ifo Institute based on the BACI dataset.
Figure 1: Changes in trade in manufacturing
Sector designation GER-WOLRDGER-BRICSGER-USAGER-EU
Manufacture of furniture, recycling
Manufacture of motor vehicles
Manufacture of ofﬁce machinery
Machinery and Equipment
Metal production and processing
Glass, ceramics
Rubber and Plastics
Chemical products
Coking, petroleum processing
Paper, publishing and printing
Wood and wood products
Leather and leather products
Textiles and wearing apparel
Food products and tobacco processing
Mining and quarrying
Agriculture and forestry, ﬁshing 5.4 1.0 5.4 4.8
9.8 12.0 12.4 12.9
7.3 5.2 6.3 7.1
1.1 2.3 11.8 2.8
2.7 0.1 11.9 4.7
6.0 5.4 13.5 6.6
6.6 3.4 12.5 6.7
15.3 17.5 17.8 15.5
9.6 8.2 12.8 9.5
8.1 7.4 15.5 8.6
4.4 5.4 14.6 5.4
10.1 8.0 17.4 10.3
8.0 5.1 16.1 8.7
6.7 4.4 18.9 7.6
8.4 6.4 18.6 8.6
6.0 3.3 13.7 6.8

12
3.3 Regional Data
For the analysis at the regional level, we utilize data from the German Federal Statistical Office. Here
we use each state’s data segmented by economic sectors for exports to the United States from the
year 201211, as well as employment figures from the manufacturing industry from the year 200812.
This data enables us to derive conclusions on sector-based trade and value creation effects in every
state.
3.4 Company Data
The interconnection of trade effects and employment information necessary for our analysis is
made through the Linked-Employer-Employee Dataset (LIAB) of the IAB. More specifically, we
use the LIAB Cross-section Dataset 2, in which information on all employees registered for social
security was added to a sample of establishments between 1993 and 2010.13 This dataset thus
includes not only detailed information about personal characteristics of individual employees, but
also important information about the company. Of particular interest for our analysis is information
on the level of a company’s international activity. This linking of company and personnel data
allows us to estimate the effects of a potential trade agreement on wages and the income risk facing
employees. The LIAB data, however, are a stratified sample of companies. The use of weighting
factors enables us to make representative statements about the distribution of companies in
Germany.
3.5 Wage data
To be able to make representative statements about the distribution of education levels and
occupational groups in the specific sectors, we use the SIAB dataset (sample of the integrated
labor market biographies) of the IAB.14 This dataset is a representative 2 percent sample of all
employment subject to social security obligations in the Federal Republic of Germany. It contains
the most important human capital characteristics of employees and makes it possible to calculate
(daily) wages.15 These wages, inflation-corrected by the Consumer Price Index, are the basis of our
11 Statitisches Bundesamt (https://www-gene-sis.destatis.de/genesis/online/data;jsessionid=83BF49DC2CDF812C2DE72CE895
6C3355.tomcat_GO_1_2?operation=abruftabelleAbrufenselectionname=51000-0036levelindex=1levelid=1379954988568
index=5)
12 Statitisches Bundesamt (https://www-gene-sis.destatis.de/genesis/online/data;jsessionid=83BF49DC2CDF812C2DE72CE895
6C3355.tomcat_GO_1_2?operation=previouslevelindex=3levelid=1379955209838levelid=1379955194721step=2)
13 The data basis comes from the cross-section model (Version 2, 1993–2010) of the Linked-Employer-Employee data of the IAB.
Access to the data was achieved during visits to the Research Data Center of the Federal Agency for Labor at the Institute for
Labor Market and Occupational Research (FDZ). For more information, see Heining, Scholz, Seth (2013).
14 For more information, see Berge, König, Seth (2013).
15 The dataset has several well-known weaknesses. The most important is that the income variable is only filled if the person
involved has an income from employment that is below the upper limit for social security contributions. For those employment
relationships where this does not apply, there are algorithms for imputing it that have proven themselves in the literature.
3. Data and trends

13
3. Data and trends
calculations of the real wage effects of a TTIP. We also use wage data to draw conclusions about
effects of a TTIP on income disparity. Figure 2 illustrates changes in wage disparity in Germany
in selected segments of the economy, represented by the standard deviation of the logarithm of
wages. As already explained above, this is an appropriate measure of the extent of the unequal
distribution of wage income.
The economic sectors shown offer only slight differences as to the extent of the measured disparity.
The disparity is higher in the food industry than in machinery or automotive manufacturing, but
in the period average of the sectors, the measurements generally scatter around 0.35, which also
roughly applies for the German economy as a whole.16 Across a large portion of the sectors, the
trend is rising: inequality rose significantly from 1998 to 2007.
16 See for example Dustmann et al. (2009), Card et al. (2012) and Baumgarten (2012).
Source: Calculations by the ifo Institute on the basis of the SIAB dataset of the IAB Nürnberg.
Figure 2:Wage disparity in Germany
Recycling
Motor vehicles manufacturing Ofﬁce machineryMachinery equipment
Metal production and processingGlass, ceramics
Rubber and plasticsChemicals
Textiles and wearing apparelAgriculture and forestry, ﬁshing
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998
0.25
0.30
0.35
0.40
0.45
2007200420011998

14
3. Data and trends
Figure 3 shows the extent of the wage disparity that cannot be explained by the characteristics of
the employee, such as education, work experience, gender, immigrant status, etc. That is to say,
this is a residual of a Mincer equation.17
This residual can also be interpreted as the significance
of factors that are not directly under the employee’s influence. These include which company
the employee works for, but there is also the simple influence of chance (e. g., that the personnel
manager gives a salary bonus for reasons that an external user of the data cannot account for).
Figure 3 makes it clear that most of the level of inequality and changes shown in Figure 2 are due
to the residual portion of the wage inequality. In other words, it is not changing human capital
characteristics that explain Figure 2. This makes it clear that changes in the implicit price of
human capital or participation of employees in company success may lag behind development.
The theoretical and empirical literature shows very clearly that international trade impacts income
distribution specifically through these channels.18
17 Separate Mincer equations were calculated for each sector; within the sectors, the data was pooled across the years, however.
18 See Felbermayr et al. (2013) and Baumgarten (2012).
Source: Calculations of the ifo Institute based on the SIAB dataset from IAB Nürnberg.
Figure 3: Residual wage disparity in Germany
Recycling
Motor vehicles manufacturing Ofﬁce machineryMachinery equipment
Metal production and processingGlass, ceramics
Rubber and plasticsChemicals
Textiles and wearing apparelAgriculture and forestry, ﬁshing
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998
0.15
0.25
0.35
2007200420011998

15
4. How are the trade effects distributed across industries?
4. How are the trade effects distributed across
industries?
4.1 Trade creation and NTB quantification
We are now ready to conduct the first step of the research approach sketched in section 2. We
start by quantifying the expected trade effects in the 16 manufacturing industries. As already
mentioned in Part 1 of our study, the trade effect anticipated from a TTIP is the one that can
actually be observed in the data from existing trade agreements. This effect, as we will see in
greater detail below, comes not from the elimination of customs duties but mainly from lowering
non-tariff barriers (NTBs). We assume symmetry, i. e., that imports and exports are similarly
affected in terms of how they change. The effects documented are, as already emphasized in the
introduction, partially analytical in nature. They are related only to the expected change in trade
between Germany and the USA and represent its lower limits because the endogenous adaptation
of gross domestic products (which in turn has a trade-increasing effect) has not been taken into
consideration.19
The third column in Table 1 shows the changes in bilateral trade expected from a possible
transatlantic trade and investment partnership. It is evident that besides the food and tobacco
processing industries, the metal industry would profit from such an agreement. There we expect
trade growth of more than 50 percent. A similarly strong increase in the trade flow of just under
50 percent can also be seen for agriculture and forest products. Also evident is that for these
sectors in particular, the reduction of non-tariff barriers will play a decisive role. Likewise, the
manufacture of office machinery and data processing equipment will clearly benefit from a trade
agreement between the EU and USA. There the expected increases are in the 40 percent range.
For the chemical industry as well as for the furniture manufacturing sector, the expected trade
growth is around 20–26 percent.
19 In particular, the reported effects do not necessarily add up to the effects occurring in the whole economic equilibrium; see Part
1 of the study.

16
Moderate, but definitely positive trade creation effects are moreover to be expected in leather,
paper and printing, glass and machinery. Even Germany’s characteristic auto sector profits from
a transatlantic agreement. Here we expect growth between 15 and 20 percent. Increases in auto
production are relatively smaller than in other sectors because the trade relationships are already
at a comparatively high level. We don’t expect any positive growth in trade in the textile and
clothing sector, on the other hand. For mining, forestry, coking and petroleum as well as glass,
no statistically significant trade effects can be demonstrated in existing agreements. The other
columns in Table 1 indicate the average customs duties that the USA or EU charge on their imports
in the specific sectors. They are high for clothing in both regions; the EU charges significantly
higher duties for food and autos than the USA; the USA charges higher duties than the EU in
the coking/petroleum and leather sectors. Overall, the tariff rates are relatively low. To help in
understanding the process, we would add here that the tariff rates indicated for all trading partners
of the EU or the USA apply to the extent that they are members of the World Trade Organization
(WTO). It is assumed that the TTIP would result in a complete elimination of these duties.
Table 1: Sector trade effects of TTIP on EU-US trade, underlying decline in tariff and non-tariff barriers
NACE
Rev.1.1
Sector designation Trade creation
(in percent)*
Trade elasticities
according to Broda
Weinstein (QJE
2006)
Customs
Importer USA
from EU
(in percent)**
Customs
Importer USA
from EU
(in percent)**
NTB Importer
USA from EU
(in percent)**
NTB Importer
USA from EU
(in percent)**
A B Agriculture and forestry, fishing and fish
farming
47.40 1.33 2.62 3.89 33.02 31.75
C Mining and quarrying . 5.32 0.96 0.77 . .
DA Food products and tobacco processing 65.86 3.65 2.31 5.60 15.74 12.45
DB Textiles and wearing apparel –19.35 1.89 7.00 8.19 . .
DC Leather and leather products 17.35 0.96 7.10 3.91 10.97 14.16
DD Wood and wood products . 0.83 0.19 0.96 . .
DE Paper, publishing and printing 14.68 1.55 0.02 0.02 9.45 9.45
DF Coking, petroleum processing . 3.36 6.63 1.50 0.00 0.00
DG Chemical products 21.65 3.75 1.71 1.86 4.06 3.91
DH Rubber and Plastics 14.80 1.34 1.71 1.86 9.33 9.18
DI Non-metallic mineral products, ceramics . 1.30 2.56 3.11 . .
DJ Basic metals and fabricated metals 52.65 2.77 1.67 1.66 17.34 17.35
DK Machinery and Equipment 16.42 1.10 1.26 1.25 13.66 13.67
DL Manufacture of office machinery, data proces-
sing equipment and installations
39.93 2.74 0.58 0.35 13.99 14.22
DM Auto makers 16.88 2.27 1.19 4.67 6.25 2.77
DN Furniture, jewelry, musical instruments, sports
equipment, recycling Recycling
26.36 0.55 0.84 0.96 47.10 46.98
* „.“ means that econometrically, no effect could be identified that was significantly different from zero.
** Source of customs data:TRAINS Data from WITS.The customs are import-weighted average.
Source: Calculations by the ifo Institute on the basis of BACI data.

17
The table also presents the trade elasticities calculated by Broda and Weinstein (2006), which have
been aggregated here at the sector level. The higher they are, the more strongly trade between the
countries reacts to changes in trade costs. Based on this information, it is now possible to quantify
at the sector level what drop in non-tariff barriers combined with the assumed elimination of
tariffs would generate the calculated trade effects. Based on the different average duties applied,
these differ only slightly between the EU and the USA.
The last two columns in Table 1 show the calculated NTB index as ad valorem equivalents. That
means the values should be read as percentage surcharges on the manufacturing price and
thus be interpreted similar to customs duties, but with one basic difference: They are not in fact
customs duties. It turns out that the meaning of non-tariff barriers and the potential for lowering
them through bilateral agreements like a TTIP in individual industries varies. Adjustable NTBs are
especially high in furniture making or agriculture and forestry. The latter sectors especially include
many protectionist measures that claim to be protecting consumers from harmful foodstuffs from
abroad. But in machinery, metals and food, the non-tariff barriers are also substantial. Lower
adjustable NTBs can be identified for the auto and chemicals industries.
The NTBs reported describe cost savings in the non-tariff area that have already been achieved
on average by existing trade agreements. It can be assumed that a TTIP would be neither more
nor less successful. Let us again be clear: The values should not be understood as levels, but
as the likely changes in NTBs. In fact, the NTB potentials calculated can be regarded as lower
limits, because in the existing trade agreements, there is typically no complete use of the tariff
elimination potential made by the trading companies.20
4.2 Indirect effects on the service sectors
The preceding analysis has made it clear which manufacturing sectors can expect an increase in
trade from a TTIP. Now we will consider how this economic revitalization, through inter and intra-
industry links, affects Germany’s whole economy but especially its service sector. We quantify
these effects with the help of input-output analysis.21 This provides us with information on how
dependent the production of a sector is on intermediate products from its own and other sectors of
the economy. In this way it is possible to quantify indirect effects of an increase in trade in sectors
that are not themselves directly affected by a TTIP, for which, due to reasons of data availability and
quality, no direct effects can be calculated. Table 2 shows the results. The third column shows the
induced trade volumes we expect in the 16 manufacturing sectors from a transatlantic agreement.
Column four shows the direct production effects. These represent how much production in Sector
j is changed by the trade impact in Sector j. It should be noted that the induced trade volume is
20 This is due to the fact that the bureaucratic hurdles to making products duty-free are high (especially presentation of a certificate
of origin).
21 The basis is the World Input Output Dataset (WIOD) from 2007. The WIOD is available for a period from 1995–2009. We base
our calculations on 2007 in order to avoid distortions from the financial and economic crisis in 2009. To ensure a degree of
consistency, we therefore refer throughout this report to 2007, when we are using the WIOD.

18
only slightly different than the direct production effects. This is due precisely to the prior work
that Sector j requires from its own industry for manufacturing its final products. Since no direct
trade effect can be calculated for the service sectors, there is no direct production effect in these
sectors. The fifth column shows the total production effect that can be expected in Sector j. This
total effect consists of the direct production effect and the indirect production effect. The indirect
effect is the total of previous work in Sector j used to produce the final products in all other sectors
(except j). In the service sector, this indirect effect is simultaneously the total effect. In order to be
able to estimate the amplitude of the effects, column six shows the total production of the specific
sector from 2007 and in the last column, the share of the total production effect in total production.
What can be seen is that the induced production effects are in the range of up to two percent
of total production in 2007. The relatively largest production increases are to be expected in the
industries that produce office equipment. The metal products and chemicals industries are also
expected to realize significant production increases from TTIP.
The service sector most strongly affected by indirect effects is leasing movables without operators
or the sector that provides services mainly to companies. Here it is evident that the indirectly
induced effects alone add up to a share of 0.5 percent in total production of the sector.
Table 2 thus makes it very clear that although the direct trade effects occur exclusively in the
16 manufacturing sectors, the total economy would be noticeably affected by a transatlantic
agreement. We find overall that the service sector experiences a comparable impact from indirect
effects to the manufacturing sector.22
22 It is worth remembering that this is a partial analytical model, allowing abstraction from general equilibrium effects as well as
from trade diversion effects. Such effects are explained in the first part of our study (Part 1: Macroeconomic Effects).

19
Table 2: Quantification of direct and indirect production effects of a TTIP in Germany by sector
NACE
Rev.1.1
Sector designation induced
trade
creation (in
millions of
Euros)
direct
production
effect (in
millions of
Euros)
total produc-
tion effect
(in millions
of Euros)
total pro-
duction (in
millions of
Euros)
total
production
effect / total
production
A B Agriculture and forestry, fishing 102 106 230 51,950 0.004
C Mining and quarrying 0 0 33 13,710 0.002
DA Food products and tobacco processing 609 680 745 154,120 0.005
DB Textiles and wearing apparel –61 –61 –55 23,730 –0.002
DC Leather and leather products 47 47 47 3,480 0.014
DD Wood and wood products (without furniture production) 0 0 75 25,540 0.003
DE Paper, publishing and printing 416 474 633 88,900 0.007
DF Manufacture of coke, refined petroleum products and nuclear fuel 0 0 150 61,050 0.002
DG Chemical products 2,863 2,937 3,040 156,970 0.019
DH Rubber and Plastics 185 194 423 65,690 0.006
DI Non-metallic mineral products, ceramics 0 0 122 41,090 0.003
DJ Basic metals and fabricated metals 3,474 4,068 4,887 230,160 0.021
DK Machinery and Equipment 2,158 2,401 2,677 224,800 0.012
DL Manufacture of office machinery, data processing equipment and installa-
tions, manufacture of electrical and optical equipment
4,506 4,883 5,154 209,390 0.025
DM Manufacture of motor vehicles 3,538 4,210 4,337 350,720 0.012
DN Manufacture of furniture, jewelry, musical instruments, furniture, sports
equipment, toys and other instruments Recycling
288 302 369 39,160 0.009
G-50 Retail trade (not including motor vehicles or service stations); Repairs of
personal and household goods
0 0 182 54,360 0.003
G-51 Wholesale trade and commission trade (not including motor vehicles) 0 0 742 170,260 0.004
G-52 Sale; Maintenance and repair of motor vehicles and personal and
household goods
0 0 659 146,230 0.005
H Hotels and restaurants 0 0 13 66,500 0.000
I-60 Land transport services; Pipeline transport services 0 0 312 69,350 0.004
I-61 Water transport 0 0 23 25,080 0.001
I-62 Air transport 0 0 71 27,370 0.003
I-63 Other supporting and auxiliary transport activities; Activities of Travel
Agencies
0 0 354 96,580 0.004
I-64 Post and telecommunications 0 0 186 81,100 0.002
J Financial intermediation 0 0 468 219,710 0.002
K-70 Real estate activities 0 0 573 328,350 0.002
K-71-
74
Renting of machinery and equipment without operator, data processing,
research and development, provision of services predominantly for
enterprises
0 0 2321 442,530 0.005
L Public administration, defense, compulsory social security 0 0 94 182,560 0.001
M Education 0 0 74 122,380 0.001
N Health and social work 0 0 7 221,320 0.000
O Other community, social and personal services 0 0 355 167,180 0.002
P Households as employers 0 0 0 7,070 0.000
Source: Calculations by ifo Institute based on WIOD 2007.

20
4.3 Value creation and employment effects
We now show how much value creation can be expected from the induced trade effects and how
many jobs in the specific sectors would be affected by it. Here again, we distinguish between direct
and indirect effects.
Let us first examine the effects on value creation in manufacturing. The greatest total value creation
effect is recorded in the office equipment and electronics manufacturing sector. As column five in
Table 3 clearly shows, the relative growth in value creation (measured as a share of the total value
creation effect of the sector in 2007) for this sector, at 2.5 percent, is the highest. Similarly strong
value creation effects can be seen in metal production and chemicals. In the service sector, the
value creation effects again come from indirectly induced effects. The greatest increase is posted
by movables leasing, with a relative increase of 0.5 percent of value creation.
A similar picture can be obtained from examining the employment effects. Here too the sectors
already mentioned show the greatest effects. The total for all sectors results in a total employment
effect of nearly 160,000 employees. This value is only slightly different from the value of 181,000
jobs created reported by us in the macro study. The difference can be explained by the different
model structure (partial analysis versus general equilibrium) and the different datasets used.

21
Table 3:Value creation and employment effects in Germany by sector
NACE
Rev.1.1
Sector designation Direct value
creation effect
(in millions of
Euros)
Total value
creation effect
(in millions of
Euros)
Total value
creation effect
/total value
creation 2007
Direct
effect
workers
Total
effect
workers
Total worker
effect/all
workers 2007
A B Agriculture and forestry, fishing 43 93 0.004 911 1,967 0.002
C Mining and quarrying 0 12 0.002 0 197 0.002
DA Food products and tobacco processing 166 182 0.005 3,785 4,145 0.004
DB Textiles and wearing apparel –20 –17 –0.002 –393 –352 –0.002
DC Leather and leather products 13 13 0.014 297 300 0.013
DD Wood and wood products (without furniture production) 0 22 0.003 0 402 0.003
DE Paper, publishing and printing 175 234 0.007 3,039 4,060 0.007
DF Manufacture of coke, refined petroleum products and
nuclear fuel
0 11 0.002 0 49 0.002
DG Chemical products 986 1,021 0.019 8,494 8,794 0.019
DH Rubber and Plastics 69 150 0.006 1,163 2,540 0.006
DI Non-metallic mineral products, ceramics 0 46 0.003 0 692 0.003
DJ Basic metals and fabricated metals 1,297 1,558 0.021 18,790 22,570 0.020
DK Machinery and Equipment 882 983 0.012 11,597 12,932 0.012
DL Manufacture of office machinery, data processing
equipment and installations, manufacture of electrical and
optical equipment
1,817 1,917 0.025 23,204 24,490 0.024
DM Manufacture of motor vehicles 1,065 1,097 0.012 11,786 12,143 0.012
DN Manufacture of furniture, jewelry, musical instruments,
furniture, sports equipment, toys and other instruments,
recycling
102 125 0.009 1,998 2,439 0.008
G-50 Retail trade (not including motor vehicles or service
stations); repairs of personal and household goods
0 121 0.003 0 2,776 0.003
G-51 Wholesale trade and commission trade (not including
motor vehicles)
0 433 0.004 0 6,207 0.004
G-52 Sale; Maintenance and repair of motor vehicles and
personal and household goods
0 370 0.005 0 13,163 0.004
H Hotels and restaurants 0 7 0.000 0 282 0.000
I-60 Land transport services; Pipeline transport services 0 145 0.004 0 3,932 0.004
I-61 Water transport 0 6 0.001 0 20 0.001
I-62 Air transport 0 19 0.003 0 165 0.003
I-63 Other supporting and auxiliary transport activities;
Activities of Travel Agencies
0 142 0.004 0 2,129 0.003
I-64 Post and telecommunications 0 89 0.002 0 1,168 0.002
J Financial intermediation 0 184 0.002 0 2,248 0.002
K-70 Real estate activities 0 458 0.002 0 679 0.001
K-71-
74
Renting of machinery and equipment without operator,
data processing, research and development, provision of
services predominantly for enterprises
0 1,517 0.005 0 23,022 0.004
L Public administration, defense, compulsory social security 0 64 0.001 0 1,363 0.001
M Education 0 57 0.001 0 1,335 0.001
N Health and social work 0 5 0.000 0 122 0.000
O Other community, social and personal services 0 214 0.002 0 3,575 0.002
P Households as employers 0 0 0.000 0 0 0.000
Source: Calculations by the ifo Institute based on WIOD 2007,

22
5. Impact of education and occupational groups and impact of regions
5. Impact of education and occupational groups and
impact of regions
Now that we have quantified the projected trade affects and their indirect effects, we can turn
to analysis step 2 and transform the trade shocks into shocks for the various education and
occupational groups and regional shocks.
5.1 Descriptive analysis of the education groups
Within the framework of our analysis we distinguish three groups with different education
levels: first, relatively unskilled workers, i. e., those who have no vocational training and nothing
equivalent to a high school diploma. Workers with moderate qualifications have either a high
school diploma or completed an apprenticeship, while highly qualified workers have completed a
technical training college or a university degree.23
Table 4 shows the distribution of the education groups in manufacturing sectors. This shows the
shares of the different education groups in total employment in the sectors and the trade creation
potential calculated in the first part of our analysis. We examine the 16 manufacturing sectors for
which we can calculate both direct and indirect effects. As discussed earlier, direct effects cannot
be calculated for the service sector.
A more exact impression of how important the specific sectors are for each education group is
provided by Tables 5 through 7, which list the individual groups for each of the five most important
sectors. The first value in Table 5 should be understood as showing that 17 percent of all unskilled
workers in the manufacturing sector are employed in the metal production and processing industry.
23 This classification is standard in the literature and follows such examples as Dustmann et al. (2009) and Baumgarten (2012). It
should be noted here that the uncleaned education variable in the IAB dataset is of relatively poor quality. That means that many
entries are missing or that individuals have inconsistent entries. Before our analysis, we clean the education variable using the
imputation procedures that are recognized in the literature. See Fitzenberger et al. (2006) on this point.

23
Table 4: Distribution of education groups across manufacturing sectors
NACE
Rev.1.1
Selected sectors Unskilled
(in %)
Moderately
skilled (in %)
Highly skilled
(in %)
Trade creation
(in %)
A B Agriculture and forestry, fishing 19.43 75.52 5.05 47.40
C Bergbau und Gewinnung von Steinen 15.81 76.18 8.01 .
DA Food products and tobacco processing 18.96 76.98 4.06 65.86
DB Textiles and wearing apparel 16.75 75.91 7.34 –19.35
DC Leather and leather products 20.14 73.51 6.34 17.35
DD Wood and wood products 16.75 75.91 7.34 .
DE Paper, publishing and printing 20.14 73.51 6.34 14.68
DF Coking, petroleum processing 16.11 78.68 5.21 .
DG Chemical products 18.94 70.36 10.70 21.65
DH Rubber and plastics 17.87 74.08 8.05 14.80
DI Glass, ceramics 13.88 78.10 8.02 .
DJ Metal production and processing 16.95 76.10 6.95 52.65
DK Machinery and equipment 11.37 74.92 13.71 16.42
DL Manufacture of office machinery 10.90 69.06 20.04 39.93
DM Manufacture of motor vehicles 10.00 74.13 15.87 16.88
DN Manufacture of furniture, recycling 13.88 78.51 7.61 26.36
Source: Calculations of the ifo Institute based on the IAB SIAB dataset.
Table 5: Significance of specific sectors for unskilled workers
Ranking Sector designation Relative significance for unskilled workers (in %)
1 Metal production and processing 17.08
2 Food products and tobacco processing 15.26
3 Machinery and equipment 10.36
4 Paper, publishing and printing 9.98
5 Manufacture of office machinery 9.47
Table 6: Significance of specific sectors for moderately skilled workers
Ranking Sector designation Relative significance for moderately skilled workers (in %)
3 Food products and tobacco processing 12.16
5 Manufacture of motor vehicles 9.30
Table 7: Significance of specific sectors for highly skilled workers
Ranking Sector designation Relative significance for highly skilled workers (in %)
3 Manufacture of motor vehicles 13.25
5 Chemical products 8.04

24
Figure 4 shows in each case the correlation between the induced trade effects and the sector share
of the individual education groups. It illustrates that for all education groups, there is a positive
link between trade creation in a sector and the relative meaning of this sector for the specific
group. It is striking, however, that the correlation for highly skilled workers at 0.17 is noticeably
lower than for the other two groups. This is mainly due to the fact that highly skilled workers play
a subordinate role in the sectors that show especially strong trade effects (such as food products,
metal production and agriculture).
Quelle: Berechnungen des ifo Institutes auf Basis des SIAB Datensatzes des IAB Nürnberg.
0.00 0.05 0.10 0.15 0.20 0.25
–0.3
–0.2
–0.1
–0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.00 0.05 0.10 0.15 0.20 0.25
–0.3
–0.2
–0.1
–0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.00 0.05 0.10 0.15 0.20 0.25
–0.3
–0.2
–0.1
–0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Tradecreation
Moderately skilled Highly skilledUnskilled
Tradecreation
Tradecreation
TEX TEX TEX
LFF
LED
CHEM
NAHR
MTL
BURO
RCY
GUM PVD
KFZ
MB
LEDLED
GUM
CHEM
RCY
CHEM
RCY
KFZ
MB
PVD
LFF
NAHR
MTL
BURO
LFF
NAHR
MTL
BURO
KFZ
MB
GUM PVD
Figure 4: Correlation of trade creation in specific sectors
LFF = Agriculture and forestry, fishing
LED = Leather and leather products
CHEM = Chemicals
MB = Machinery and equipment
BERG = Mining and quarrying
HOLZ = Wood and wood products
KFZ = Motor vehicles
TEX = Textiles and wearing apparel
OEL = Coking, petroleum processing
MTL = Metal production and processing
RCY = Furniture production, recycling
GUM = Rubber
BURO = Office machinery
NAHR = Food products and tobacco processing
PVD = Paper, publishing and printing
GLAS = Glass, ceramics
Sector shareSector share Sector share

25
5.2 Impact of education groups
The trade shock from analysis step 1 we now transform into impact measurements. While in the
first part we identified the impact on specific sectors from a potential TTIP, we are now trying
to quantify the impact on specific education groups. We are following the method proposed by
Ebenstein et al. (2012) and calculate the index in two steps. First, we quantify a weighting factor
2010kjα :
(1)
2010
2010
2010
kj
kj
k
L
L
α = ,
Where 2010kjL is the number of employees in education group k in sector j in 2010, and 2010kjL
the number of all employees in education group k in all sectors in 2010. Using the weighting factor,
we then calculate the impact measurement. 2010kβ :
(2) 2010 2010
1
J
k kj j
j
β α
=
= Δ∑
,
in which j∆ represents the trade creation effect in sector j calculated in analysis step 1.
Table 8 shows the susceptibility index for the three education groups. This confirms again
the impression that the correlation analysis has already provided: Unskilled workers are most
significantly affected by the direct trade shock in the manufacturing sectors. Then come the
moderately skilled workers and least affected are the highly skilled workers.
5.3 Descriptive analysis of the occupational groups
The IAB dataset distinguishes more than 300 different occupations.24 To obtain a degree of clarity,
we have combined the individual listings into a total of 88 occupational groups.25 Table 9 provides
an initial impression of which occupational groups are in the sectors that show the greatest trade
creation potential. A value of 34 % for sales personnel in the first part of the table thus means that
34 % of all employees in the food sector belong to this occupational group.
24 The occupational classification in the IAB dataset follows the one used by the Bundesagentur für Arbeit 1988 and includes some
330 characteristics.
25 We aggregate the employment categories (3 places), into employment groups (2 places).
Table 8: Education impact measurement
Impact measurement education
Unskilled 0.34
Moderately skilled 0.31
Highly skilled 0.27

26
5.4 Impact on occupational groups
Here we again calculate the impact, this time on individual occupational groups. Tables 9 i and 9
ii in the Appendix show the impact measurements of all 88 occupational groups in detail. What
is clear is that those occupations that work almost exclusively in the food production industry are
most likely to be affected. These include, for example, sales personnel. The metal industry is also
affected more than average. At the lower end of this scale are occupational groups like printers,
paper makers and construction materials suppliers and occupations in the clothing industry.
However, it becomes clear that some of the service industries are affected rather more heavily.
Here for example is where you find hospitality services workers (index of 0.50) or cleaners (index
of 0.33). This shows once more that the total economy is affected by a TTIP, even if in this section,
we only look at the direct trade effects on manufacturing employment.
5.5 Descriptive analysis of the regions
In this section, we analyze how much the individual states in Germany would be affected by
possible transatlantic trade and investment partnership. Table 10 shows the two most important
sectors per state with regard to their exports to the USA in the year 2012. We see, for example,
that in Bavaria, auto manufacturing and the electronics sector generate the greatest percentage of
exports. In Lower Saxony, it is machinery and auto manufacturing. An additional table (Table 10i)
in the Annex shows which states have the greatest share of total German exports to the USA per
sector. Here we see, for example, that for agriculture, Lower Saxony is the state with the highest
export percentage in trade with the USA, followed by Bavaria and Schleswig-Holstein.
Table 9:The most important occupational groups in certain sectors
Food products and tobacco
processing
Metal production and
processing
Agriculture and forestry,
fishing
Manufacture of office
machinery
Manufacture of furniture,
recycling
Sales personnel
(34 %)
Metalworkers
(15 %)
Horticulturalists
(36 %)
Office personnel and assistants
(14 %)
Carpenters
(26 %)
Producers of bakery and pastry
products (11 %)
Assembly workers and metal-
workers (12 %)
Agricultural workers
(26 %)
Electricians
(10 %)
(14 %)
Meat and fish processors
(9 %)
Office personnel and
assistants (11 %)
(5 %)
Assembly workers and metal-
workers (9 %)
Laborers
(6 %)
Other food occupations
(6 %)
Metal workers
(6 %)
Farmers
(4 %)
Engineers
(9 %)
Warehouse, transport workers
(5 %)
Office personnel and
assistants (6 %)
Laborers
(5 %)
Forestry and hunting occupa-
tions (3 %)
Technicians
(8 %)
Wood workers, makers of
wood products (4 %)

27
Table 10:The most important industrial sectors by state
State
Baden-Württemberg
43.7 % Automotive manufacturing
22.8 % Machinery
15.1 % Office machinery manufacturing
Bavaria
15.9 % Machinery
Berlin
19.6 % Machinery
Brandenburg
58.0 % Chemicals
Bremen
10.3 % Food and tobacco processing
4.3 % Metal production and processing
Hamburg
8.6 % Machinery
Hesse
43.8 % Chemicals
11.2 % Machinery
Mecklenburg-Vorpommern
32.8 % Machinery
16. 5% Office machinery manufacturing
Lower Saxony
9.8 % Machinery
9.8 % Chemicals
North Rhine-Westphalia
25.5 % Machinery
23.1 % Chemicals
Rhineland-Palatinate
60.9 % Chemicals
12.9 % Machinery
Saarland
24.2 % Machinery
Saxony
14.7% Machinery
Saxony-Anhalt
40.9% Chemicals
14.0 % Machinery
Schleswig-Holstein
29.8 % Machinery
27.7 % Chemicals
Thuringia
26.3 % Machinery
Source: Calculations of the ifo Institute for Economic Research based on data from the German Federal Statistical Office

28
5.6 Impact on regions
Based on that data, how would the calculated trade effects from a TTIP be distributed among the
states? If we assume that the trade effects on individual industries are distributed equally among
the states, the following results emerge (see Table 11):
Accordingly, we expect overall that import increases in bilateral trade with the USA could reach
20 to 30 percent per state. North Rhine-Westphalia could increase its exports to the United States
by about 29 percent, due primarily to the strong position of metal production and processing in
that state.
We expect the least impact in Saxony and Brandenburg, two regions whose exports to the USA
are limited to a few sectors, such as the chemicals sector in Brandenburg. These sectors have a
comparatively low forecast for trade creation.
The relatively strong values for Mecklenburg-Vorpommern, Thuringia and Bremen are mainly due
to strong trade-creating effects in the food production industry and its importance in these states.
Table 11: Expected export increases by state
State
Mecklenburg-Vorpommern 29 %
North Rhine-Westphalia 29 %
Thuringia 28 %
Berlin 27 %
Hesse 26 %
Saxony-Anhalt 26 %
Schleswig-Holstein 25 %
Saarland 25 %
Bremen 24 %
Rhineland-Palatinate 24 %
Lower Saxony 23 %
Bavaria 23 %
Baden-Württemberg 22 %
Hamburg 22 %
Brandenburg 21 %
Saxony 20 %
Source: Calculations of the ifo Institute on the basis of the LIAB dataset of the IAB.

29
5.7 Value creation and employment effects in the regions
In addition to the bilateral export effects, we can also draw conclusions about the regional
employment market and value creation effects. However, the calculation is limited to the direct
trade effects on the manufacturing industry. In light of this, we expect that the actual extent of the
effects will be substantially higher, since around 40 percent of the newly created jobs would fall
under the non-exportable service sector.
Table 12 demonstrates that three states in particular would benefit from a free trade agreement:
North Rhine-Westphalia, Baden-Württemberg and Bavaria. This is due primarily to their already
high export levels. Approximately 60 percent of all manufacturing exports to the USA come from
these states. In addition, a larger number of manufacturing industries that would see the greatest
value creation effects from an agreement are located there, especially machinery manufacturing,
metal production and processing, electronics industry (office machinery manufacturing, etc.), and
auto manufacturing. And finally, Bavaria and Baden-Württemberg alone account for almost 20
percent of Germany’s exports to the USA.
Table 12: Regional impact on the manufacturing industry
State Total employment growth Total value creation effect
(in millions of Euros)
North Rhine-Westphalia 21,080 1,433
Baden-Württemberg 20,163 1,566
Bavaria 19,471 1,597
Lower Saxony 7,647 555
Hesse 6,796 599
Rhineland-Palatinate 4,500 425
Saxony 4,014 207
Thuringia 2,477 101
Schleswig-Holstein 2,116 146
Saxony-Anhalt 1,986 72
Berlin 1,722 145
Saarland 1,460 100
Brandenburg 1,452 158
Hamburg 1,198 121
Mecklenburg-Vorpommern 735 25
Bremen 551 199
Source: Calculations of the ifo Institute on the basis of the LIAB dataset of the IAB.

30
5.8 Employment and education in the regions
In this section, we take a look at how the calculated employment effects are distributed among the
different education groups in each state. To do so, we combine the findings from section 5.1 with
those from section 5.7 and derive the corresponding distribution for the individual states.
Table 13 illustrates the corresponding results for all states and education groups. For Germany’s 16
states, it shows essentially no major deviations from the expected average values.
However, a few noteworthy differences emerged between states in the areas of relatively unskilled
and highly qualified workers.
For example, employment growth among relatively unskilled workers in Rhineland-Palatinate was
highest in the national comparison.
The obvious reason for this effect is that the chemical sector, which employs a comparatively high
percentage of relatively unskilled workers, is particularly important in this state.
Table 13: Employment effects by state based on education level
State Employment
growth for
relatively
unskilled workers
Share of
relatively
unskilled workers
in employment
growth
Employment
growth for
workers with
moderate
qualifications
Share of workers
with moderate
qualifications
in employment
growth
Employment
growth for highly
qualified workers
Share of highly
qualified workers
in employment
growth
Total employ-
ment growth
Baden-Württemberg 2,708 13.4 % 14,761 73.2 % 2,694 13.4 % 20,163
Bavaria 2,632 13.5 % 14,199 72.9 % 2,640 13.6 % 19,471
Berlin 243 14.1 % 1,234 71.6 % 246 14.3 % 1,722
Brandenburg 219 15.0 % 1,073 73.9 % 160 11.0 % 1,452
Bremen 67 12.2 % 408 74.0 % 76 13.8 % 551
Hamburg 153 12.8 % 871 72.7 % 174 14.5 % 1,198
Hesse 984 14.5 % 4,943 72.7 % 869 12.8 % 6,796
Mecklenburg-Vorpommern 104 14.2 % 542 73.7 % 89 12.1 % 735
Lower Saxony 1,087 14.2 % 5,635 73.7 % 925 12.1 % 7,647
North Rhine-Westphalia 3,159 15.0 % 15,567 73.8 % 2,354 11.2 % 21,080
Rhineland-Palatinate 709 15.8 % 3,305 73.4 % 486 10.8 % 4,500
Saarland 212 14.5 % 1,095 75.0 % 153 10.5 % 1,460
Saxony 570 14.2 % 2,951 73.5 % 493 12.3 % 4,014
Saxony-Anhalt 310 15.6 % 1,465 73.8 % 211 10.6 % 1,986
Schleswig-Holstein 309 14.6 % 1,548 73.2 % 258 12.2 % 2,116
Thuringia 362 14.6 % 1,813 73.2 % 301 12.2 % 2,477
[ Manufacturing industry [ 16.6 % [ 75.1 % [ 8.8 %
Total 13,827 14.2 % 71,411 73.3 % 12,129 12.5 % 97,368
Quelle: Berechnungen des ifo Institutes auf Basis der Daten des statistischen Bundesamtes.

31
Interestingly, the direct comparison of the total potential employment effects from an agreement
with the current existing average distribution of the three education groups in the manufacturing
industry shows that the relative increases for highly qualified workers are significantly greater
than for the other two groups (see Table 13). At 12.5 percent, the growth here lies almost 4
percentage points higher than the current sectoral average of 8.8 percent for highly qualified
workers. By contrast, the growth for moderately skilled and relatively unskilled workers came
in at around 2 percent below the current sector ratio. This can be explained by the fact that in
sectors such as machinery manufacturing and electronics in particular, where most jobs in the
manufacturing industry are created, the percentage of highly qualified workers is especially high
compared to the other sectors. In the electronics industry alone, it amounts to almost 20 percent.
At this point we would like to remind readers again that the expected employment distribution
calculated here only takes the manufacturing industry into account, and therefore does not include
the service sector due to a lack of data.

32
6. Effects of a TTIP on income and the income risk
We are now turning to our third analysis step and use the calculated shock measurement to project
changes in real income and in the income risk that we would expect as the result of a TTIP.
6.1 Effects on real income
For this we first examine Mincer wage equations. Such equations model the wages of dependent
employees as a function of their characteristics. Typically for this type of analysis, individual wage
data (in logarithmic form) are regressed on human capital indicators (education, work experience),
socio-economic variables (age, gender, nationality), and a series of indicator variables (region,
industry). In this way, information is obtained on the role of education in the wages paid, for
example.
For our analysis, we expand the classic model by including characteristics of the employer. What
interests us in particular is to what extent the establishment where the worker is employed is
affected by international trade (exports and imports).
Concretely, we estimate the following Mincer regressions on individual wage data for 2010::
(3) ln i i s iw Xβ γ ε= + Ω + ,
where ln iw represents the logarithm of real wages, iX is a vector of variables controlling for the
characteristics of the worker, and β
the related parameter vector.26
sΩ on the other hand, measures
how strongly an establishment is involved in international trade; we view these variables as an
openness measurement.27
Accordingly, the estimated value of parameter γ
provides information
about the strength of the link between the openness of the establishment and the wages of the
employees.
26 Since the wage information in the IAB dataset is available only to the upper earnings limit, we first use Tobit estimates, in order
to establish the complete wage distribution from graded wage data. In doing this, we follow Dustmann et al. (2009) and Card et al.
(2012). For our analysis we are using the imputed wage data. Moreover, we take into account the disproportionality of the LIAB
calculation sample in that we additionally control for the industry, state and size of operation variables. For more information,
see FDZ Method Report No. 01/2008.
27 Concretely, the openness measurement measures the extent of exports in the total sales of a company. Since the extent of
exports at the company level correlates strongly with the extent of imports at the company level (see on this e. g. Baumgarten,
2012), this variable can be used as a proxy for the total openness of a company. We consider the average as the openness of a
sector.

33
We conducted separate regressions for each group and thereby obtained group-specific γ
values.
They offer insight into how strongly openness influences the real wages in the specific sample.
Table 14 shows the results of the Mincer regressions for the individual education groups. The
corresponding results for all occupational groups can be found in the Appendix as Tables 14 i.
The estimates shown in the Table should be interpreted as follows: A 10 % increase in openness in
a sector (and accordingly, it is assumed, in an average establishment) leads to an average increase
in real wages of 4 % for unskilled and 3 % for moderately and highly skilled workers. This finding
is somewhat surprising. It means that the export success of individual companies also results in
higher wages for unskilled workers and that the increase in wages in export-oriented companies
is actually higher for less skilled persons than for workers with more training.
In further conducting our analysis, we use the shock measurements calculated previously in the
estimated Mincer equations. In so doing, we take into account that the openness approximated is
only with respect to trade by German firms with the USA. This means that openness, for example,
in the food products sector rises by 66 % * 7 %, with 7 % representing the ratio of Germany’s
exports to the USA relative to Germany’s total exports.28 We thereby obtain a forecast of real wage
change '
ˆγ
, which should result from a TTIP on average for the individual groups. Table 15 shows
the forecast of real wage change for individual education groups.
28 We are assuming this 7 % value for 2010 as an average value for all sectors.
Table 14: Regression results for education groups
Variables Unskilled Moderately skilled Highly skilled
Individual characteristics yes yes yes
Openness 0,004*** 0,003*** 0,003***
Number of observations 358.768 1.739.263 310.905
R² 0,4964 0,2943 0,3377
*** indicates statistical significance at 1%.
Source: Calculations by the ifo Institute based on the IAB LIAB dataset.
Table 15: Forecasts of real wages
Unskilled Moderately skilled Highly skilled
TTIP effect on real wages 0.0094 0.0065 0.0056
Source: Calculations of the ifo Institute on the basis of the IAB LIAB dataset.

34
Different aspects should be noted here. First, it is evident that all three values are positive.
That means that in case of a TTIP, we expect real wage increases in all three education groups.
For unskilled workers, we are forecasting an increase in real wages of about 0.9 percent, for
moderately skilled an increase of 0.7 percent and for highly skilled workers we expect an increase
in real wages of 0.6 percent.29
For the individual occupational groups, we are forecasting real wage changes that range between
minus 5 percent (fishery jobs) and plus 5 percent (agricultural workers). For sales personnel,
inspectors and dispatchers, builders of civil engineering structures and precision metal workers,
we identify real wage increases of more than one percent. That also applies for meal preparers,
social workers and personal care workers. This again makes clear that the whole economy would
profit substantially from a transatlantic agreement. Because even though in this second segment
we are looking only at the direct trade effects on manufacturing, it can be seen that for example
cleaners or other service workers obtain real wage increases.
6.2 Effects on income risk
The effects of a TTIP on income risk are calculated in a similar way. As measure for this we use the
residual wage disparity, i. e., the share of the wage inequality that cannot be ascribed to individual
characteristics like age, education, gender or nationality. It thus represents a risk measurement,
since it cannot be controlled for an individual.30 If the scatter of these random wage components
grows due to an increase in trade, then the income risk on the labor market increases: In other
words, there is an increase in the share of workers paid substantially more or less than the wage
their formal qualifications would entitle them to.
The procedure takes place in three steps. First, we conduct Mincer wage regressions and
extract the residuals from them. The standard deviation of these residuals provides our analysis
measurement, the income risk. In the second step, we estimate the extent the income risk is
influenced by openness. To do that, we regress our income risk index on three constructed openness
measurements, so that we can reach some conclusions about individual subsegments of the job
market. In the last step, we again use the impact measurements that we had calculated to see how
the income risk for various partial segments changes as the result of a possible agreement.31
First we look at the isolated results for the three education groups. Our analysis shows that for all
29 These results are compatible with the real wage effects for Germany that we calculated in the first part of the study (Part 1:
Macro Economic Effects). There the real wage effects were between 0.5 and 2.2 percent, where we distinguish between a purely
customs scanerio and a deep liberalization scenario. That distinction is not made here. Moreover, our analysis does not consider
that the price level in the whole economy falls due to TTIP (as we know from the macro study), so that the real wage effects
shown here represent lower limits. If the price level effect is assumed to be equal for all population groups (which is the usual
practice), then the real wage effect can be readily compared across individual groups.
30 In econometric estimation equations, the unexplained residual is usually described as “shocks”. These scatter around zero in a
linear estimation model. In our application, a lower residual means that the worker receives a higher wage, which would be due
to him based on his calculated human capital profile. A positive residual means the opposite.
31 The construction of the openness index is again based on the plant information of the LIAB database. The first openness index
varies across education groups, the second varies across occupational groups and the third openness index varies across states.

35
education groups, an increase in the income risk in the case of a transatlantic investment and trade
partnership would be expected. In fact, the risk for unskilled workers rises most, with the standard
deviation rising by 0.011. Workers with moderate skills can expect an increase in the income risk
of 0.010, while for highly skilled workers, the income risk rises by 0.0089. These results should
be interpreted against a background in which the average residual wage inequality has a value of
0.4.32 We should thus see the average value as the status quo and then expect that a TTIP would
increase the income risk of unskilled workers by nearly three percent. The corresponding income
risk rise for the other two education groups is just below that.
A different picture emerges if we look at the isolated effects for individual occupational groups.
Here it turns out that the occupations more frequently found in sectors that are distinguished
for a high openness measurement, are subject to a smaller income risk. (That means that our
openness factor in analysis step two described above is negative for the occupational groups). The
result: For almost every occupational group, there is a decline in the income risk from a possible
TTIP. The income risk will decline more, the more the occupational group is impacted by trade
or openness. Table 16 i in the Annex provides the relevant overview. The effects from a TTIP on
income risk on various occupational groups now lies between plus 0.0015 (textile processors) and
minus 0.0058 (meat and fish processors). Again, the more open an occupational group is (i. e., the
more important open sectors are for this occupational group), the more the income risk for this
occupational group falls from a TTIP.
So far we have examined the effects on the income risk for education and occupational groups
in isolation. To reach some conclusions about the overall effect, we must evaluate the total of the
findings. Although there are opposite effects, the overall effect of a TTIP on the income risk is
ambiguous. What we can say, however, is for example, that for an unskilled (= 0.011) concrete
worker (= –0.0014), the income risk rises. For an unskilled salesperson (=0.0066), the income risk
rises, but less than for the concrete worker. Thus it is possible to calculate for combinations of the
education and occupation of a worker the expected change in his income risk.
32 This is a value that is also found in this amplitude in the literature. See Dustmann et al. (2009), Card et al. (2012) and Baumgarten (2012).

36
7. Summary
7. Summary
• Trade creation potential varies across industry sectors: The strongest trade effects are to be
expected in the food and metal industries.
• There are also clear differences in the sectors with respect to the importance of non-tariff
barriers. Politically adaptable NTBs are especially high in the recycling and agriculture
sectors.
• The largest value creation effects in Germany are expected in the electrical sector. That is
where the strongest employment effects are found.
• The transatlantic agreement also has affects in those industrial sectors and on those
workers that are not directly affected by more trade. The reason for that are input-output
interconnections.
• Real wage increases for all education groups (unskilled, moderately skilled and highly skilled)
are to be expected from a TTIP, from 0.6 percent (highly qualified) to nearly one percent
(unskilled).
• Occupations in the food sector, such as sales personnel, like occupations in the metal industry,
show above average real wage increases of more than one percent.
• All states would benefit from an agreement, and increases in trade with the USA of 20–30
percent per state are expected. The size of the anticipated effects depends heavily on export
levels at the outset.
• The income risk from a TTIP rises in all education groups and regions. However, the total effect
remains ambiguous, since the income risk falls along the occupational dimension.

37
A. Annex
A. Annex
A.1 Gravitation models, estimation methods and results
The economic gravitation equation in its simplest form states that the trade flows between two
economies depend proportionally on the product of their size and negatively on their distance
from each other:
(A1) 1i j
ij ijW
BIPBIP
x d
BIP
−
=
,
where ijx
stands for the trade flows and
W
BIP for world income, and iBIP
and jBIP

correspondingly for the GDP of countries i and j. ijd
measures the trade barriers between two
countries. Although the simple gravitation equation in the empirical literature on trade is generally
able to explain 60–80 % (depending on the dataset) of the variation in bilateral trade flows, the
absence of a solid theoretical foundation for this conclusion was long considered a major criticism.
Current research has however been able to show that the gravitation equation is consistent with
many newer trade models.33
A theoretically-based gravitation equation from Anderson and van
Wincoop (2003) is very similar to the simple equation (A1):
(A2)
(1 ) ( 1) ( 1)i j
ij ij i jW
BIPBIP
x d P
BIP
σ σ σ− − −
= Π ,
where σ
reflects substitutional elasticity, and iΠ
and jP
represent the multilateral resistance
terms. Trade policy is represented as a part of ijd , in that we integrate an indicator variable for
membership in a preferential trade agreement (PTA).34
We assume the following connection:
(A3)
1
exp( ...)ij ij ijd PHA DISTσ
δ β−
= + + ,
where we consider, besides the indicator for membership in a preferential trade agreement, other
geographic and historical variables that influence trade frictions between two countries. So, for
example, ijDIST stands for the distance between the trading partners. By substitution of (A3) in
(A2) and with a slight modification, we obtain our estimation equation:
(A4) ( )'
ln ln lnij ij ij i j ijx Z PHAβ δ α γ ε= + + + + ,
where
'
(1, ,...)ijZ DIST= is a vector that contains a constant as well as all variables that make
trade easier or more difficult except for ijPHA
. β
is a vector of coefficients and
1
i i iBIP σ
α −
= Π

and
1
j j jBIP Pσ
γ −
=
apply.
33 This applies only under certain assumptions involving consumption, preferences, market structure and transport costs. Decisive
contributions were made by Redding and Venabeles (2004), Baier and Bergstrand (2001), Anderson and van Wincoop (2003)
and Anderson and van Wincoop (2004).
34 In this we again distinguish between “deep” and “all other” preferential trade agreements. We include among the “deep”
agreements the North American Free Trade Agreement (NAFTA) and the EU Agreement.

38
A. Annex
For our analysis, we estimate the gravitation equation at the industry level and obtain for each
sector a coefficient sδ
that indicates to us the average trade-creating effect of a deep agreement
for that particular sector. Gravitation models can be estimated consistently with the help of
fixed effects (Feenstra, 2004). In specification A) we estimate our industry-specific gravitation
equations, in that we control for trade partner/sector-specific fixed effects, as well as for country/
time-specific fixed effects and perform a linear estimate.
The fixed effects control for observed and unobserved heterogeneity in the data. In specification
B) we conduct a non-linear, so-called Poisson Pseudo Maximum Likelihood (PPML) estimate, in
order to take into account such factors as the relatively high number of zeros in the trade data
(Santos Silva and Tenreyro, 2006). The PPML estimate has another decisive advantage, besides the
consideration of the zeros in the trade matrix, compared to linear estimation methods: It generates
consistent estimators even in the presence of measurement errors that cause heteroscedasticity.
In B) time-consistent components are included through trade partner/sector-specific fixed effects.
Instead of country/time-specific fixed effects that include the multilateral resistance terms in
specification A), in specification B) we take into account time-specific fixed effects and linear
multilateral resistance terms. In this we make use of the findings of Baier and Bergstrand (2009)
and approximate the multilateral resistance terms with help of a Taylor approximation.

39
A. Annex
Table 1 i shows the results of both specifications. Together the results of both models – both
in terms of the sign and the scale of the effects – are comparable. Since the trade effects at the
sector level are the foundation for further calculations, Table 1 i also presents how they result from
specification A and B.
Note to Table 1 i:
The estimates shown also prove to be robust in alternative specifications.
Table 1 i: Estimation results
NACE
Rev.1.1
Sector designation Specification A Specification B Derived trade creation
A B Agriculture and forestry, fishing 0,728*** 0,388*** 0,388***
C Mining and quarrying –0,0249 0,0774 0,0774
DA Food products and tobacco processing 0,714*** 0,506*** 0,506***
DB Textiles and wearing apparel –0,0106 –0,215** –0,215**
DC Leather and leather products 0,160* –0,140 0,160*
DD Wood and wood products 0,0448 –0,0675 –0,0675
DE Paper, publishing and printing 0,221*** 0,137* 0,137*
DF Coking, petroleum processing 0,00668 0,133 0,133
DG Chemical products 0,300*** 0,196** 0,196**
DH Rubber and Plastics 0,138** 0,105 0,138**
DI Glass, ceramics 0,0337 0,0300 0,0300
DJ Metal production and processing 0,423*** 0,0325 0,423***
DK Machinery and Equipment 0,0971 0,152** 0,152**
DL Manufacture of office machinery 0,0734 0,336*** 0,336***
DM Manufacture of motor vehicles 0,159* 0,156* 0,156*
DN Manufacture of furniture, recycling –0,00709 0,234** 0,234**
*, **, *** refer to significance at the 1.5 or 10 percent level. Shortened representation: Since we estimate each sector
separately, we will not provide references to additional information here.
Source: Calculations of the ifo Institute.

40
A. Annex
Table 9 i: Degree of impact experienced by the occupational groups
Occupational group designation Impact measurement
Farmers 0.43
Livestock breeders, fishing occupations 0.43
Administrators, farming and livestock consultants 0.39
Agricultural workers, animal keepers 0.47
Horticulturalists 0.47
Forestry, hunting occupations 0.46
Miners .
Mineral, petroleum, natural gas extractors .
Mineral processors .
Stone workers 0.01
Construction materials producer .
Ceramic workers 0.02
Glass makers 0.07
Chemical workers 0.21
Plastic processors 0.18
Paper makers, processors 0.14
Printers 0.15
Wood processors, wood product producers and related occupations 0.08
Metal processors, rollers 0.48
Form makers, casters 0.45
Metal formers (die casters) 0.39
Metal formers (under tension) 0.31
Metal surface processors, enhancers, coaters 0.42
Metal connectors 0.34
Blacksmiths 0.42
Sheet metal worker, fitters 0.24
Metal workers 0.31
Mechanics 0.27
Tool makers 0.30
Precision metal worker and associated occupations 0.37
Electrician 0.28
Installer and metal working occupations not named elsewhere 0.33
Spinning occupations –0.12
Textile producers –0.10
Textile processors –0.10
Textile finishers –0.17
Leather producers, leather and pelt processors 0.18
Producers of baked goods and pastries 0.65
Meat and fish processors 0.65
Meal preparers 0.55
Food and drink producers 0.62
Other food occupations 0.65
Masons, concrete installers 0.16
Carpenter, roofer, scaffolder 0.13
Road and foundation workers 0.29
Construction laborer 0.32

41
A. Annex
Table 9 i: Degree of impact experienced by the occupational groups
Occupational group designation Impact measurement
Construction materials supplier 0.14
Interior designer, upholsterer 0.20
Cabinetmaker, model building 0.18
Painters, varnishers related occupations 0.26
Goods testers, dispatching packers 0.34
Laborers without more detailed designation 0.31
Machinists and related occupations 0.26
Engineers 0.26
Chemists, physicists, mathematicians 0.22
Technicians 0.26
Technical support personnel 0.27
Sales personnel 0.50
Bank, insurance sales personnel 0.05
Other service sales personnel and related occupations 0.22
Land transport occupations 0.37
Water and air transport occupations 0.21
Telecommunications occupations 0.21
Warehouse managers, warehouse, transport workers 0.23
Entrepreneurs, organizers, auditors 0.26
Elected officials, administrative decision-makers 0.25
Accountants, IT specialists 0.28
Office staff, clerical workers 0.26
Service, security occupations 0.30
Security personnel 0.07
Journalists, interpreters, librarians 0.15
Artists and associated occupations 0.21
Doctors, pharmacists 0.18
Other health service providers 0.18
Social work professions 0.24
Teachers 0.16
Intellectual and scientific occupations not mentioned elsewhere 0.22
Personal care 0.30
Hospitality service providers 0.51
Domestic services providers 0.33
Cleaning occupations 0.36
Workers with unidentified occupations 0.27
Workers without more detailed designation 0.29
Source: Calculations of the ifo Institute based on the IAB LIAB dataset.

42
A. Annex
Tabelle 10 i:The most important states by industrial sector
Sector designation State
Agriculture and forestry, fishing
27.9% Lower Saxony
24.0% Bavaria
17.4% Schleswig-Holstein
Mining and quarrying
30.2% Bavaria
26.9% Saxony-Anhalt
12.9% Lower Saxony
Food products and tobacco processing
24.1% Bremen
20.9% Lower Saxony
11.1% North Rhine-Westphalia
Textiles and wearing apparel
26.2% Baden-Württemberg
24.5% Bavaria
Leather and leather products
28.8% Rhineland-Palatinate
25.9% Bavaria
20.5% Schleswig-Holstein
19.4% Bavaria
13.8% Lower Saxony
29.0% Lower Saxony
Coking, petroleum processing
59.9% Hamburg
Chemical products
18.8% Hesse
Rubber and Plastics
17.9% Bavaria
16.0% Hesse
Glass, ceramics
36.4% Bavaria
Metal production and processing
10.2% Hesse
Manufacture of office machinery
33.8% Bavaria
Machinery and Equipment
21.5% Bavaria
Manufacture of motor vehicles
30.2% Bavaria
11.6% Lower Saxony
Manufacture of furniture, recycling
22.4% Bavaria
15.4% Hesse
Source: Calculations of the ifo Institute.

43
A. Annex
Table 14 i: Regression results, occupational groups,TTIP effect
Occupational group designation Openness TTIP effect
Farmers 0.005 0.0000
Livestock breeders, fishing occupations –0.016** –0.0479
Administrators, farming and livestock consultants 0.005* 0.0136
Agricultural workers, animal keepers 0.015*** 0.0495
Horticulturalists 0.002 0.0000
Forestry, hunting occupations 0.005 0.0000
Miners 0.001 .
Mineral, petroleum, natural gas extractors –0.007*** .
Mineral processors 0.001* .
Stone workers 0.003* 0.0002
Construction materials producer 0.003*** .
Ceramicists 0.001* 0.0001
Glass makers –0.002 0.0000
Chemical workers 0.002** 0.0030
Plastic processors 0.001** 0.0013
Paper makers, processors 0.001 0.0000
Printers 0.000 0.0000
Wood processors, wood product producers and related occupations 0.002 0.0000
Metal processors, rollers 0.002*** 0.0068
Form makers, casters 0.001 0.0000
Metal formers (die casters) 0.001 0.0000
Metal formers (under tension) 0.003*** 0.0065
Metal surface processors, enhancers, coaters 0.002** 0.0059
Metal connectors 0.002*** 0.0048
Blacksmiths 0.002 0.0000
Sheet metal worker, fitters 0.003*** 0.0050
Metal workers 0.003*** 0.0065
Mechanics 0.002*** 0.0038
Tool makers 0.002*** 0.0043
Precision metal worker and associated occupations 0.004*** 0.0104
Electrician 0.002*** 0.0040
Installer and metal working occupations not named elsewhere 0.002*** 0.0046
Spinning occupations 0.001 0.0000
Textile producers 0.002 0.0000
Textile processors 0.001 0.0000
Textile finishers 0.003*** –0.0035
Leather producers, leather and pelt processors 0.001 0.0000
Producers of baked goods and pastries 0.003 0.0000
Meat and fish processors –0.000 0.0000
Meal preparers 0.004*** 0.0155
Food and drink producers 0.002 0.0000
Other food occupations 0.000 0.0000
Masons, concrete installers 0.002** 0.0023

44
A. Annex
Table 14 i: Regression results, occupational groups,TTIP effect
Occupational group designation Openness TTIP effect
Carpenters, roofers, scaffolders 0.001 0.0000
Road and foundation workers 0.006** 0.0120
Construction laborer 0.002 0.0000
Construction materials supplier 0.003** 0.0029
Interior designer, upholsterer –0.003 0.0000
Cabinetmaker, model building 0.004*** 0.0050
Painters, varnishers related occupations 0.004*** 0.0072
Goods testers, dispatching packers 0.005*** 0.0117
Laborers without more detailed designation 0.007*** 0.0154
Machinists and related occupations 0.002*** 0.0036
Engineers 0.002*** 0.0037
Chemists, physicists, mathematicians 0.001** 0.0015
Technicians 0.002*** 0.0036
Technical support personnel 0.002*** 0.0038
Sales personnel 0.009*** 0.0317
Bank services, insurance sales personnel 0.003* 0.0011
Other service sales personnel and related occupations 0.004*** 0.0061
Land transport occupations 0.001 0.0000
Water and air transport occupations 0.004** 0.0060
Telecommunications occupations 0.003*** 0.0045
Warehouse managers, warehouse, transport workers 0.004*** 0.0066
Entrepreneurs, organizers, auditors 0.001 0.0000
Elected officials, administrative decision-makers 0.003*** 0.0052
Accountants, IT specialists 0.004*** 0.0080
Office staff, clerical workers 0.003*** 0.0055
Service, security occupations 0.004*** 0.0084
Security personnel 0.001 0.0000
Journalists, interpreters, librarians –0.001** –0.0010
Artists and associated occupations 0.005*** 0.0074
Doctors, pharmacists 0.005*** 0.0063
Other health service providers 0.003* 0.0039
Social work professions 0.009*** 0.0152
Teachers 0.006*** 0.0066
Intellectual and scientific occupations not mentioned elsewhere 0.004*** 0.0063
Personal care 0.020* 0.0424
Hospitality service providers 0.002* 0.0071
Domestic services providers 0.001 0.0000
Cleaning occupations 0.004*** 0.0100
*, **, *** indicates significance level at 1.5 and 10 percent level.

45
A. Annex
Table 16 i: Effect on income risk by occupation
Occupational group designation Impact measurement TTIP effect on income risk
Farmers 0.43 –0.0038
Livestock breeders, fishing occupations 0.43 –0.0038
Administrators, farming and livestock consultants 0.39 –0.0034
Agricultural workers, animal keepers 0.47 –0.0042
Horticulturalists 0.47 –0.0041
Forestry, hunting occupations 0.46 –0.0041
Miners . .
Mineral, petroleum, natural gas extractors . .
Mineral processors . .
Stone workers 0.01 –0.0001
Construction materials producer . .
Ceramicists 0.02 –0.0002
Glass makers 0.07 –0.0006
Chemical workers 0.21 –0.0019
Plastic processors 0.18 –0.0016
Paper makers, processors 0.14 –0.0013
Printers 0.15 –0.0013
Wood processors, wood product producers and related occupations 0.08 –0.0007
Metal processors, rollers 0.48 –0.0043
Form makers, casters 0.45 –0.0039
Metal formers (die casters) 0.39 –0.0034
Metal formers (under tension) 0.31 –0.0027
Metal surface processors, enhancers, coaters 0.42 –0.0037
Metal connectors 0.34 –0.0030
Blacksmiths 0.42 –0.0037
Sheet metal worker, fitters 0.24 –0.0021
Metal workers 0.31 –0.0027
Mechanics 0.27 –0.0024
Tool makers 0.30 –0.0027
Precision metal worker and associated occupations 0.37 –0.0033
Electrician 0.28 –0.0025
Installer and metal working occupations not named elsewhere 0.33 –0.0029
Spinning occupations –0.12 0.0011
Textile producers –0.10 0.0009
Textile processors –0.10 0.0009
Textile finishers –0.17 0.0015
Leather producers, leather and pelt processors 0.18 –0.0016
Producers of baked goods and pastries 0.65 –0.0058
Meat and fish processors 0.65 –0.0058
Meal preparers 0.55 –0.0049
Food and drink producers 0.62 –0.0055
Other food occupations 0.65 –0.0057
Masons, concrete installers 0.16 –0.0014

46
A. Annex
Table 16 i: Effect on income risk by occupation
Occupational group designation Impact measurement TTIP effect on income risk
Carpenters, roofers, scaffolders 0.13 –0.0011
Road and foundation workers 0.29 –0.0025
Construction laborer 0.32 –0.0029
Construction materials supplier 0.14 –0.0012
Interior designer, upholsterer 0.20 –0.0018
Cabinetmaker, model building 0.18 –0.0016
Painters, varnishers, related occupations 0.26 –0.0023
Goods testers, dispatching packers 0.34 –0.0030
Laborers without more detailed designation 0.31 –0.0028
Machinists and related occupations 0.26 –0.0023
Engineers 0.26 –0.0023
Chemists, physicists, mathematicians 0.22 –0.0019
Technicians 0.26 –0.0023
Technical support personnel 0.27 –0.0024
Sales personnel 0.50 –0.0044
Bank, insurance sales personnel 0.05 –0.0005
Other service sales personnel and related occupations 0.22 –0.0019
Land transport occupations 0.37 –0.0032
Water and air transport occupations 0.21 –0.0019
Telecommunications occupations 0.21 –0.0019
Warehouse managers, warehouse, transport workers 0.23 –0.0021
Entrepreneurs, organizers, auditors 0.26 –0.0023
Elected officials, administrative decision-makers 0.25 –0.0022
Accountants, IT specialists 0.28 –0.0025
Office staff, clerical workers 0.26 –0.0023
Service, security occupations 0.30 –0.0027
Security personnel 0.07 –0.0006
Journalists, interpreters, librarians 0.15 –0.0013
Artists and associated occupations 0.21 –0.0019
Doctors, pharmacists 0.18 –0.0016
Other health service providers 0.18 –0.0016
Social work professions 0.24 –0.0021
Teachers 0.16 –0.0014
Intellectual and scientific occupations not mentioned elsewhere 0.22 –0.0020
Personal care 0.30 –0.0027
Hospitality service providers 0.51 –0.0045
Domestic services providers 0.33 –0.0029
Cleaning occupations 0.36 –0.0031
Workers with unidentified occupations 0.27 –0.0024
Workers without more detailed designation 0.29 –0.0026

48
Literatur
Literature
Anderson, James, and Eric van Wincoop. “Gravity with Gravitas: A Solution to the Border
Puzzle”. American Economic Review (93) 1 2003. 170–192.
Anderson, James, and Eric van Wincoop. “Trade Costs”. Journal of Economic Literature (42) 3
2004. 691–751.
Baumgarten, Daniel. “Exporters and the rise in wage inequality: Evidence from German linked
employer-employee data”. Journal of International Economics (90) 1 2013. 201–217.
Baier, Scott and Jeffrey Bergstrand. “The Growth of World Trade: Tariffs, Transport Costs, and
Income Similarity”. Journal of International Economics (53) 1 2001. 1–27.
Baier, Scott and Jeffrey Bergstrand. “Bonus vetus OLS: A Simple Method for Approximating
International Trade-Cost Effects using the Gravity Equation”. Journal of International
Economics (77) 1 2001. 77–85.
Vom Berge, Philipp, Marion König and Stefan Seth. “Stichprobe der Integrierten Arbeitsmarkt-
biografien (SIAB) 1975–2010”. FDZ Datenreport 1 2013.
Broda, Christian, and David Weinstein. “Globalization and the Gains from Variety”. Quarterly
Journal of Economics (121) 2 2006. 541–585.
Card, David, Jörg Heining and Patrick Kline. “Workplace Heterogeneity and the Rise of West
German Wage Inequality”. The Quarterly Journal of Economics (128) 3 2013. 967–1015.
Dustmann, Christian, Johannes Ludsteck and Uta Schönberg. “Revisting the German wage
structure”. Quarterly Journal of Economics (124) 2 2009. 843–881.
Ebenstein, Avraham, Ann Harrison, Margret McMillan and Shannon Philipps. “Why are
American Workers getting Poorer? Estimating the Impact of Trade and Offshoring Using the
CPS” NBER Working Papers 15107, National Bureau of Economic Research, Inc 2009.
Feenstra, Robert. Advanced International Trade: Theory and Evidence. Princeton University
Press, Princeton NJ 2004.
Felbermayr, Gabriel, and Erdal Yalcin. “Export Credit Guarantees and Export Performance: An
Empirical Analysis for Germany.” The World Economy, forthcoming (doi: 10.1111/twec.12031).

49
Literature
Felbermayr, Gabriel, Andreas Hauptmann and Hand-Jörg Schmerer. “International Trade and
Collective Bargaining Outcomes: Evidence from German Employer-Employee Data”.
Scandinavian Journal of Economics, forthcoming.
Felbermayr, Gabriel, Mario Larch, Lisandra Flach, Erdal Yalcin and Sebastian Benz. Dimensionen
und Auswirkungen eines Freihandelsabkommens zwischen der EU und den USA. ifo Studie.
München 2013.
Fischer, Gabriele, Florian Janik, Dana Müller and Alexandra Schmucker. “Das IAB Betriebspanel –
von der Stichprobe über die Erhebung bis zur Hochrechnung”. FDZ Methodenreport 1 2008.
Fitzenberger, Bernd, Aderonke Osikominu and Robert Völter. “Imputation rules to improve
the education variable in the IAB employment subsample”. Schmollers Jahrbuch: journal of
Apllied Social Science Studies (126) 3 2006. 405–436.
Heining, Jörg, Theresa Scholz and Stefan Seth. “Linked-Employer-Employee-Daten des IAB: LIAB-
Querschnittsmodell 2 1993–2010”. FDZ Datenreport 2 2013.
Redding, Stephen and Anthony Venables. “Economic Geography and International Inequality”.
Journal of International Economics (62) 1 2004. 53–82.
Santos Silva, Joao, and Silvana Tenreyro. “The Log of Gravity”. Review of Economics and Statistics
(88) 4 2006. 641–658.

50
Global Economic Dynamics (GED)
About the Authors
Prof. Gabriel J. Felbermayr, Ph. D., Director of the Ifo Center for International Economics
ifo Institut – Leibniz Institute for Economic Research at the University of Munich
Sybille Lehwald
Dr. Ulrich Schoof
Project Global Economic Dynamics (GED), Bertelsmann Stiftung, Gütersloh.
Mirko Ronge
Project Global Economic Dynamics (GED), Bertelsmann Stiftung, Gütersloh.
About the Project “Global Economic Dynamics” (GED)
The Bertelsmann Stiftung established the project “Global Economic Dynamics” (GED) to shed
more light on the growing complexity of international economic relationships. By using state-of-
the-art tools and methods for measuring, forecasting and displaying the dynamics of the world
economy, the project aims at making globalization, its economic effects as well as its political
consequences more transparent and tangible.
Contact
Bertelsmann Stiftung
GED-Team
Program Shaping Sustainable Economies
Carl-Bertelsmann-Straße 256
33311 Gütersloh | Germany
Phone +49 5241 81-81353
Fax +49 5241 81-681353
ged@bertelsmann-stiftung.de
GED-Team
Program Director
Andreas Esche
Director Shaping Sustainable Economies
Phone +49 5241 81-81333
Fax +49 5241 81-681333
andreas.esche@bertelsmann-stiftung.de

51
Global Economic Dynamics (GED)
Project Managers
Dr. Jan Arpe
Project Manager
Phone +49 5241 81-81157
Fax +49 5241 81-681157
jan.arpe@bertelsmann-stiftung.de
Samuel George
Project Manager
Phone +49 5241 81-81661
Fax +1 202 384-1984
samuel.george@bfna.org
Dr. Thieß Petersen
Senior Expert
Phone +49 5241 81-81218
Fax +49 5241 81-681218
thiess.petersen@bertelsmann-stiftung.de
Dr. Ulrich Schoof
Project Manager
Phone +49 5241 81-81384
Fax +49 5241 81-681384
ulrich.schoof@bertelsmann-stiftung.de
Co-operation Partner
Poschingerstraße 5
81679 München
Contact
Prof. Gabriel J. Felbermayr, Ph. D.
Phone +49 89 9224 1428 | felbermayr@ifo.de | www.cesifo-group.de/felbermayr-g
Sybille Lehwald
Phone +49 89 9224 1250 | lehwald@ifo.de | www.cesifo-group.de/lehwald-s

54
Imprint
Imprint
© 2013 Bertelsmann Stiftung
33311 Gütersloh
Germany
www.bertelsmann-stiftung.de
Responsible
Dr. Ulrich Schoof
Autors
Prof. Gabriel J. Felbermayr, Ph. D.
Sybille Lehwald
Dr. Ulrich Schoof
Mirko Ronge
Translation
Peritus Precision Translations, Inc.
San Carlos, California, USA
www.peritusls.com
Design
Nicole Meyerholz, Bielefeld
Design Title
Markus Diekmann, Bielefeld
Picture
© koya979/Shutterstock Images

www.bertelsmann-stiftung.de
Address | Contact
33311 Gütersloh
GED-Team
Program Shaping Sustainable Economies
Phone +49 5241 81-81353
ged@bertelsmann-stiftung.de
www.ged-project.de

Study: Who Will Benefit in Germany from a TTIP?

More Related Content

What's hot (20)

Viewers also liked (18)

Similar to Study: Who Will Benefit in Germany from a TTIP? (20)

More from Bertelsmann Stiftung (20)

Recently uploaded (20)

Study: Who Will Benefit in Germany from a TTIP?