Export promotion: evaluating the impact on aggregate exports and GDP University of Geneva and International Trade Center ETPO meeting, Milan - October 14-16 2015
What do we know? Rose (2007): embassy presence increase exports by 10% Lederman et al. (2010): a 10% increase in TPO budgets increases exports by 0.6 to 1%
Problems Cross-section Endogeneity Only average returns
Unanswered questions Heterogeneity in returns by TPO? What explains that some agencies do better than others? What about GDP per capita returns?
What we did Use panel data to control for unobserved heterogeneity Second WB survey in Fall 2010 New ETPO data for 14 countries: AUT, BEL, BIH, CHE, CYP, DNK, EST, GBR, ISL, ITA, LTU, MLT, NLD, SVN Explore what type of funding, activity and structure works better using varying coefficient models Provide returns by agency as a function of characteristics
The WB surveys Contacted 149 countries in 2005 and 2010, 116 had an EPA and we sent our survey by email 94 TPOs answered, 88 positively All survey data was collected over a 4 month period All responses were followed up with phone calls to double check and clarify answers Survey contained 19 questions over structure, financing, mission, activities and expenditure
The ETPO survey Same survey but directed only towards 20+ ETPO in 2015 14 answered positively All survey data was collected over a 3 month period All responses were followed up with emails and phone calls to double check and clarify answers
Example of survey question
Descriptive statistics Share of export promotion budgets on exports Share spent by activity Share spent by firm size Share spent on established, occasional and non-exporters Public versus private sources of funding Correlation between exports and TPO budgets
Budget to export ratio in ETPO countries
Budget in export support services vs. marketing
Budget spent by firm size
Budget spent by type of exporter
Private vs public funding
Export promotion budgets and exports
Econometric results Average results Varying returns by agency characteristic Returns to exports and GDP per capita GDP per capita returns and the level of development
Average export returns 10% increase in TPOs budgets increases exports by 0.7% The 95% confidence interval is given by [0.3% - 1.2%] 1 USD increase in TPOs budgets increases exports by 87 USD The 95% confidence interval is given by [37 USD - 139 USD]
Average GDP returns 10% increase in TPOs budgets increases GDP by 0.7%: The 95% confidence interval is given by: [0.3% - 1.0%] 1 USD increase in TPOs budgets increases GDP by 384 USD The 95% confidence interval is given by [209 USD - 555 USD]
Varying coefficients - export model
Varying coefficients - GDP per capita model
What works? summary Export returns GDP returns Larger share of private sector in Executive Board Larger funding of budget from user fees Target non-exporters and established exporter Target medium and large firms Low share of spending on marketing activities Target established exporters and non-exporters (up to a max) Target small firms High share of spending in marketing activities
Export and GDP per capita returns by agency GDP growth due to TPO interventions 1 1.05 1.1 1.15 1.2 Rwanda Bangladesh Zambia Egypt, Arab Rep. Tanzania Mexico Burkina Faso Kenya Sierra Leone Nepal Senegal Ecuador Guatemala Indonesia Malawi West Bank and Gaza Cote d'ivoire Philippines Dominican Republic Peru Spain Paraguay Costa Rica Vietnam El Salvador Jamaica United Ireland Korea, Kingdom Rep. Italy Barbados Nicaragua Moldova Australia Armenia Macedonia, FYR Honduras Lebanon Austria Sweden Belize Syrian Arab Belgium Republic Yemen, Norway Botswana Netherlands Rep. Denmark Israel Albania Lithuania Switzerland Estonia Guyana Serbia Finland Bulgaria Bosnia and Herzegovina Uruguay Dominica Iceland Germany Turkey Croatia Hungary Panama Cyprus Jordan Oman Trinidad and Tobago Slovenia Malaysia Uganda Brazil 1.02 1.04 1.06 1.08 1.1 1.12 Export growth due to TPO interventions
GDP per capita returns and the level of development 1.05 1.1 1.15 1.2 Malawi Bangladesh Zambia Sierra Burkina Leone FasoKenya Nepal Rwanda Uganda Tanzania Senegal Egypt, Arab Rep. Indonesia Ecuador Guatemala West Bank and Gaza Mexico Cote d'ivoire Dominican Republic Philippines Peru Brazil Spain Paraguay Jamaica Costa Rica Korea, Rep. El Salvador Italy United Kingdom VietnamNicaragua Ireland Barbados Australia Moldova Honduras Macedonia, FYR Malaysia Armenia Austria Sweden Lebanon Syrian Arab Republic Belize Belgium Yemen, Rep. Denmark Norway Botswana Netherlands Guyana Lithuania Estonia Israel Albania Bosnia and Herzegovina Serbia FinlandSwitzerland Bulgaria Portugal Turkey Panama Hungary Dominica Uruguay Germany Iceland Croatia Jordan Cyprus Trinidad Oman and Tobago France Slovenia 4 6 8 10 12 (mean) lgdpc Malta Fitted values (mean) rgdpc
Concluding remarks Large returns in terms of exports Larger in terms of GDP per capita What works for exports may not work for GDP per capita Knowing its objective is important Need to benchmark against the objective Other characteristics may matter too