Measuring Services Trade Restrictions Aaditya Mattoo (based on research with Ingo Borchert and Batshur Gootiiz) Development Research Group The World Bank 11 March 2012
Motivation: Policy Transparency and Analysis Limited data on service trade policy, despite the importance of services sectors in terms of both output and international transactions Identifying and Quantifying Services Trade Restrictions could help in advancing national policy reform by facilitating the analysis of services policies; informing international negotiations by providing data on actual policies; 2
Structure of presentation 1. A new Services Trade Restrictions Database (available at iresearch.worldbank.org/servicestrade) 2. Measuring Services Trade Barriers 3. Do policy barriers matter? a. Investment flows b. Access to services: banking c. Trade facilitation d. Industrial policy 4. Qualifications and Looking Ahead 3
Data: country coverage 103 countries (of which 79 developing) 0.0-20.0 (37) 20.0-40.0 (44) 40.0-60.0 (17) 60.0-80.0 (4) 80.0-100.0 (1) No data (105) 4
Data: sector/mode coverage Sectors/Sub-sectors Mode 1 Mode 3 Mode 4 Banking Bank lending x x Deposit acceptance x x Insurance Automobile insurance x x Life insurance x x Reinsurance x x Telecom Fixed-line Mobile Retailing Retail distribution Transport Air passenger domestic x Air passenger international x x Maritime shipping international x x Maritime auxiliary services x Road trucking x Railway freight x Professional Services Accounting x x x Auditing x x x Legal advice foreign law x x x Legal advice dom law x x Court representation x x x x x 5
Data: policy coverage Policy dimensions covered: Barriers to foreign entry and ownership Licensing requirements Restrictions on operations Regulatory environment Focus is on measures that discriminate against foreign services and service providers; but we also cover certain non-discriminatory measures which significantly affect trade. Policy information was collected through detailed questionnaires administered by local law firms and governments were given the opportunity to comment (about half did). 6
Measuring policy restrictiveness Real value of the Database is the rich information on a range of policy variables. But there is also a need for measures that facilitate depiction of patterns and empirical analysis. Three approaches: The Services Trade Restrictiveness Index (STRI) scores based on expert judgment on a five-point scale from completely open to completely closed [0, 25, 50, 75, 100] Ordinal Ranking of Policy Combinations ranks different combinations of policy at the country-sector level in terms of relative openness Measuring restrictiveness by impact using econometric approaches estimates the restrictiveness of policies based on their impact on some outcome variable of interest, controlling for other determinants 7
Example of Ordinal Ranking: Bank Lending (mode 3) Bin 1 2 3 4 5 Example (total/bin) Argentina (38) Namibia (4) Ghana (26) India (24) Iran (1) Sector open to foreign entry Entry through establishment Operations Entry as a branch yes No restriction No restriction No restriction yes No restriction No restriction Not allowed yes No restriction Nationality req Not allowed yes Entry not allowed Nat treatment violation Nationality req; repatriation restr license limit branches n/a n/a n/a
0 20 40 60 80 100 Global patterns Country-level STRI and per-capita income: ETH ZWE IND IRN QAT DRC BDI PHL EGY IDN NPL BGD VNM JOR TUN MYS OMN BHR KWT CHN THA PAN LBN SAU LKA NAM DZA BWA MWIUGA ZAF BLR VEN TZA YEM RWA MLI KEN LSO CIVCMR NGA PAK CRIMEX UKR URY KHMUZB TUR RUS ITA FRA CHL ZMB HND MAR PRT KORJPN BEL FIN MOZ MDGGHA SEN KGZ PRY GTM ALBBRA AUS DNK CAN COL PERKAZARGHUN BGR MUS CZE GRC NICMNGBOL ROM ESP DEU AUT USA GEO ARM DOM POL LTU GBR SWE TTO NZL NLD IRL ECU 5 6 7 8 9 10 11 12 Log of GDP per capita Note: GDP per capita 2007, PPP (constant 2005 internat US$) 9
0 10 20 30 40 50 60 70 Global patterns STRI by sector and region: GCC SAR MENA EAP AFR LAC OECD ECA Note: 103 countries included. Financial Telecom Retailing Transportation Prof.Services 10
Analyzing the impact of policies 1. How do services policies shape investment flows? Cross-country, cross-sector inbound FDI flows FDI flows into banking 2. How do they affect access to services? Banking 3. Comparing quantification methodologies 11
(1) Investment flows across sectors How does policy restrictiveness affect global M&A activity? (1) (2) (3) (4) (5) (6) Log(GDP) 2007 0.6271*** 0.6466*** Log(GDP per capita) 2007 0.5958*** 0.4448*** Population Density (people/sqkm) 0.0007 0.0009* Dummy STRI (25-50] -0.3347* -1.2122*** 0.0266 Dummy STRI (50-75] -1.0844* -2.0094** 0.1350 Dummy STRI (75-100] -2.5569*** -5.1664*** -1.6993** No Investment (0%) -1.2156*** -3.0110*** 0.1062 Minority Ownership -1.2573*** -1.1678** -1.0676** Limit number of licenses 0.2131 0.4931 0.5100** Discrim licensing criteria -0.6286*** -0.8852*** -0.5587** Restriction repatr earnings -0.6951** 0.4245-0.7666* No recourse/appeal -0.6813** -1.5015*** -1.0085** Observations 744.0000 744.0000 744.0000 478.0000 478.0000 478.0000 Pseudo R^2 0.7901 0.6434 0.8792 0.7795 0.7215 0.8818 Log L -4.280e+05-7.272e+05-2.462e+05-3.613e+05-4.563e+05-1.936e+05 Sector fixed effects yes no yes yes no yes Country fixed effects no yes yes no yes yes Dependent variable is total value of sectoral M&A inflow in million USD; Poisson estimation with robust standard errors based on 8 services sectors. 12
(1) Investment flows in banking How does policy restrictiveness affect global M&A activity in banking? (1) (2) (3) Log(GDP) 2007 0.6611*** 0.6786*** 0.6544*** Log(GDP per capita) 2007 0.6133*** 0.3982*** 0.5967*** Population Density (people/sqkm) 0.0007 0.0005 0.0008 STRI (25, 50] -0.4418* STRI (50, 75] -2.5633*** STRI (75, 100] -20.9608*** Restr on Branches only 0.5189 Operational Restr -0.8227*** Major Restr Establishment -0.8571*** No acquisitions (0%) -3.5001*** Minority ownership -0.7339 Limit number of licenses -0.1728 Restriction repatriation earnings -1.0985** No recourse/appeal -0.0068 No branch entry 0.0595 Restr on cross-border trade -0.1471 Constant -1.5984 0.5918-1.3882 Observations 93 93 93 Pseudo R-sq 0.7950 0.8071 0.8016 Log L -1.158e+05-1.090e+05-1.121e+05 Dependent variable is total value of M&A inflow to the Banking sector in million USD. Poisson estimation with robust standard errors. 13
(3) Availability of credit to the private sector How does policy restrictiveness affect availability of credit in banking? (1) (2) (3) Log(GDP) 2007 28.7852*** 23.1826*** 28.9766*** Log(Population) -21.8667*** -16.3525*** -20.9216*** Population Density (people/sqkm) 0.0296* 0.0242* 0.0232 STRI (25, 50] -14.7005 STRI (50, 75] -68.1747*** STRI (75, 100] -51.2875*** Restr on Branches only -16.8306 Operational Restr -29.6466*** Major Restr Establishment -32.8754*** No acquisitions (0%) -58.1234*** Minority ownership 0.5520 Limit number of licenses -24.8711** Restriction repatriation earnings 2.2888 No recourse/appeal 2.9468 No branch entry -6.0036 Restr on cross-border trade 9.5379 Constant -12.8094 13.4692-15.7323 Observations 91 91 91 R-squared 0.4629 0.4925 0.4756 Dependent variable is domestic credit to the private sector (% GDP). Linear regression with robust standard errors. Domestic credit as a percentage of GDP is taken from the World Development Indicators for the year 2008. For Kyrgyztan the corresponding 2007 figure and for Zimbabwe the 2006 figure is used for want of missing 2008 values in both countries. 14
(3) Comparing methodologies: STRI and Econometric Estimates Estimated Index based on M&A 20 40 60 80 0 (1) (2) (3)(1) (4)(2) (3) (4) All Cntr Dev Cntr All All Cntr Cntr Dev Dev Cntr Cntr All Cntr Dev Cntr Log(GDP) 0.6240*** 2007 0.6507*** 0.6466*** 0.6240*** 0.6665*** 0.6507*** 0.6466*** 0.6665*** 7 Log(GDP 0.6052*** per capita) 2007 0.5068*** 0.4448*** 0.6052*** 0.4302*** 0.5068*** 0.4448*** 0.4302*** ple/sqkm) Population 0.0007 Density (people/sqkm) 0.0010 (1) (2) 0.0009* 0.0007 (3)(1) 0.0016** 0.0010 (4)(2) 0.0009* (3) 0.0016** (4) All Cntr Dev Cntr All All Cntr Cntr Dev Dev Cntr Cntr All Cntr Dev Cntr STRI low 0.0796-0.0305 0.0796-0.0305 Log(GDP) 0.6240*** 2007 0.6507*** 0.6466*** 0.6240*** 0.6665*** 0.6507*** 0.6466*** 0.6665*** STRI intermediate -0.3801* -0.3781-0.3801* -0.3781 7 Log(GDP 0.6052*** per capita) 2007 0.5068*** 0.4448*** 0.6052*** 0.4302*** 0.5068*** 0.4448*** 0.4302*** STRI high -1.6137** -2.3415*** -1.6137** -2.3415*** ple/sqkm) Population 0.0007 Density (people/sqkm) 0.0010 0.0009* 0.0007 0.0016** 0.0010 0.0009* 0.0016** A Comparison of the STRI with an Econometric Estimate of Restrictiveness: No Investment (0%) -1.2156*** -2.3381** -1.2156*** -2.3381** STRI low 0.0796-0.0305 0.0796-0.0305 Minority Ownership -1.2573*** -0.8826* -1.2573*** -0.8826* STRI intermediate -0.3801* -0.3781-0.3801* -0.3781 s Limit number of licenses 0.2131 0.2100 0.2131 0.2100 STRI high -1.6137** -2.3415*** -1.6137** -2.3415*** Discrim licensing criteria -0.6286*** -0.5125** -0.6286*** -0.5125** gs Restriction repatr earnings -0.6951** -0.9320** -0.6951** -0.9320** No Investment (0%) -1.2156*** -2.3381** -1.2156*** -2.3381** No recourse/appeal -0.6813** -0.5783* -0.6813** -0.5783* Minority Ownership -1.2573*** -0.8826* -1.2573*** -0.8826* s Limit number of licenses 0.2131 0.2100 0.2131 0.2100 Observations 744 520 478 744 328 520 478 328 Discrim licensing criteria -0.6286*** -0.5125** -0.6286*** -0.5125** Pseudo 0.7910 R-sq 0.7545 0.7795 0.7910 0.7406 0.7545 0.7795 0.7406 gs Restriction repatr earnings -0.6951** -0.9320** -0.6951** -0.9320** Log L -4.262e+05-1.359e+05-3.613e+05-4.262e+05-1.147e+05-1.359e+05-3.613e+05-1.147e+05 No recourse/appeal -0.6813** -0.5783* -0.6813** -0.5783* tal value of Dependent sectoral M&A variable inflow is total in million value USD; of sectoral M&A inflow in million 0 USD; 25 50 75 1 STRI (mode 3) robust standard Poisson errors estimation based with on 8 robust services standard sectors; errors based on 8 services sectors; Observations 744 520 478 744 328 520 Note: weighted average 478 of sectoral indices 328 (Table A.1) uded but not Sector reported. fixed effects included but not reported. Pseudo 0.7910 R-sq 0.7545 0.7795 0.7910 0.7406 0.7545 0.7795 0.7406 Log L -4.262e+05-1.359e+05-3.613e+05-4.262e+05-1.147e+05-1.359e+05-3.613e+05-1.147e+05 tal value of Dependent sectoral M&A variable inflow is total in million value USD; of sectoral M&A inflow in million USD; robust standard Borchert/Gootiiz/Mattoo Poisson errors estimation based with on 8 -- robust services Services standard sectors; Trade Restrictions errors based Database on 8 services sectors; 15
(3) Comparing methodologies A Comparison of the STRI with the OECD Product Market Regulation: 16
0.5 1 Services reform as trade facilitation Number of flights per airline and restrictive air transport policies: ZWE -1 -.5 PAN ETH PRY DEU MEX CANZMB NIC POL NLD DNK NZL FIN GTM HND IRL CZE PHL CRI ITAMWI BEL IDN JPN PER COL PRT CHL ECU DOM UZB GHA MOZ GBR NAM ESP KHM KOR LAO JOR CHN MAR ROM SWE VNM LTUFRA URY RWA BWA DZA QAT PAK BGD LKA USA MDG THA MYS AUT ARG YEM ALB BOL KEN IND MNG SEN TZA DRC TTO TUN LBN BHR AUS UGA UKR SAU BRA OMN MLI EGY ZAF TUR NPL MUS KWT BLR RUS ARM KGZ VEN IRN BGR GRC GEO NGA KAZ CMR -.5 0.5 1 Air Passenger STRI (high) coef = -.51037446, (robust) se =.10641397, t = -4.8 New work would find other indicators of performance and examine more rigorously the links between policy and performance Borchert/Gootiiz/Grover/Mattoo -- Landlocked or Policy Locked? 17
Services reform as industrial policy India s services reform has boosted not only productivity and exports in services, but also the performance of downstream manufacturing industries Percent Gains in Annual TFP Growth After Services Reform Study based on panel data for 4,000 Indian firms for the 1990-2005 period finds that banking, telecommunications and transport reforms all have significant positive effects on the productivity of manufacturing firms 4 3 2 1 New work would help policy-makers understand the implications of alternative sequencing of reforms in goods and services. 0 industries with limited dependence on banking industries with high dependence on banking industries with limited dependence on telecomms industries with high dependence on telecomms Source: Arnold, Javorcik, Lipscomb and Mattoo (2010).
0 20 40 60 80 100 Uruguay Round Commitments, Doha Offers and Actual Policy IND CRI GTM URY HND LKA GTM TUN TUN KEN THA PAK IDN EGY COL NGA IND NIC PHL IDN MAR LKA PRY COL MYS CHL BOL THA MUS PAN MUS PAN NGA IND DOM PAKMAR DOM BRA TUR PE R TUR PHL IDN THA MEX EGY BRA PAN MYS MEX CHN BGR ARG JOR ZAF LKA TUN ZAF PE R JOR ZAF KEN CRI URY MEX NGA CHL PAKMAR ALB TUR HND GTM PRY PECOL RBRA ARG BOL DOM NIC MUS BGR POL POL HUN HUN LTU LTUHUN POL LTU BHR TTO TTO BHR PRT KOR BHR ITA OMN OMNKOR PRT ITA GRC ESP CZE GRCITA PRT ESP CZE KOR NZL GRC ESP NZL TTO FRA DNK BEL FIN FRA SWE DEU BEL DNK JPN FIN GBR AUT CAN NLD FRAUS AUT CAN JPNDEU FINSWE BEL GBRDNK NLD AUS DEU AUT CAN JPN AUS GBR SWE NLD IRL IRL IRL USA USA 0 10000 20000 30000 40000 GDP per capita,ppp Restr ictive ness of GATS co mmitment Restr ictive ness of DOHA Offers Restr ictive ness of actual policy Fitted values Fitted values Fitted values STRI for 61 countries, excluding Qatar and 31 countries that did not submit offers Source: Borchert,Gootiiz, Mattoo 2011
Qualifications 1. Absence of cross-sectoral data on market structure means we are capturing restrictions on entry and not the prevailing extent of competition. 2. Limited information on prudential and pro-competitive regulation makes it hard to assess : how far such measures offer additional protection how far the effect of liberalization depends on the state of regulation 3. Limited information on, and difficult assessment of, actual implementation of policies, such as licensing requirements. Fortunately, some evidence in Doing Business and Investing Across Borders. 4. Paucity of comparable data on (sector-specific) outcome variables - prices, quality, diversity of services - makes it hard to infer restrictiveness of policy by estimating impact on outcomes. But we try! 20
Looking ahead Improving the coverage and quality of data Refining quantification methodologies Rigorous analyses of the impact of policy choices Investigation of the determinants of policy choices Borchert/Gootiiz/Grover/Mattoo -- Landlocked or Policy Locked? 21