Title page

Contents

Abstract 3

1/2. Evolution of FTAs 1950-2015 5

2/3. An automatization of FTAs' text analysis 5

2.1/3.1. Description of the algorithm 5

2.2/3.2. Assessment of the algorithm results 9

3/4. Business mobility clauses in trade agreements 11

4/5. Impact of facilitating business visitors entry 14

4.1/5.1. Data 14

4.2/5.2. Trade Regressions 15

4.3/5.3. Baseline results 15

4.4/5.4. Controlling for FTA's depth 19

4.5/5.5. Business travel intensity 19

4.6/5.6. Export probability 22

4.7/5.7. Permanent migration 24

5/6. Conclusion 26

Appendices 29

A. Percentage of agreements identically identified 29

B. Differences between World Bank's database and the algorithm results for the Nuclear Safety topic 31

D. PPML regressions for business trips 37

C. EU and NAFTA/CUSFTA additional controls 36

Table 1. Effect of business visitors provisions: PPML 16

Table 2. Effect of business visitors provisions: OLS 17

Table 3. Effect of business visitors provisions: PPML in shares 18

Table 4. Controlling for FTA's measures of depth 20

Table 5. Effect of business visitors provisions on Business Trips 21

Table 6. Effect of business visitors provisions on export probability 23

Table 7. Effect on business visitors provisions on migration flows (Migration Data Oecd) 24

Table 8. Effect on business visitors provisions on migration flows (WB Migration Flows) 25

Table 9. Controlling for EU/NAFTA 36

Table 10. Effect of business visitors provisions on Business Trips 37

Table 11. Effect of business visitors provisions on Business Trips 38

Figure 1. Evolution of the number of FTAs over time 6

Figure 2. Evolution of the depth of FTAs over time 7

Figure 3. Number of Agreements identified by WB vs algorithm 10

Figure 4. Percentage of FTAs coded identically for each topic 11

Figure 5. Number of FTAs signed each year containing business visitors provisions 12

Figure 6. Geographical distribution of FTAs with business visitors entry provisions 13

Figure 7. Percentage of all international trade taking place between countries with a FTA, covered by FTAs identically coded by the WB and the algorithm for... 29

Figure 8. Percentage of agreements identified by FTA coder as treating a given topic among the agreements coded by WB as treating that same topic 30

Figure 9. Percentage of agreements identified by FTA coder as not treating a given topic among the agreements coded by WB as not treating that same topic 31