Tourism Specialization and Economic Development: Evidence from the UNESCO World Heritage List (joint with Reda Cherif and John Piotrowski)

Transcription

1 Tourism Specialization and Economic Development: Evidence from the UNESCO World Heritage List (joint with Reda Cherif and John Piotrowski) BBL sponsored by the Cultural Heritage and Sustainable Tourism Thematic Group at the Urban Anchor, World Bank October 15, 2009 Rabah Arezki IMF Institute t International Monetary Fund

2 Stylized Facts Figure 1. Economic Growth and Tourism Specialization

3 Motivation Broad question: Is tourism specialization a viable development strategy? The question we tackle: What is the true magnitude of tourism specialization s impact on economic growth? Our result could be used as a benchmark to address the broad question for individual countries.

4 Channels of effect A2 In theory it could go in both directions: Tourism and Dutch Disease: Chao et al. (2006). Services and growth: Baumol (1967), Oulton (2001). Network effects, in addition to enhanced security, stability, etc. A1 Empirical results: Overall positive. Spillovers of FDI on growth: Borenztein et al. (1998). Tourism and growth using panel regressions: Sequeria and Nunes (2008).

5 Slide 4 A1 A2 How about the magnitude? compared to ours! ACHERIF, 10/13/2009 You could say that in theory we have both signs. Empricially it is slightly positive (magnitude?). However, no emprirical study corrects for potential bias given by the many theoretical channels. ACHERIF, 10/13/2009

6 Contribution ti Estimation of a standard growth model augmented with tourism specialization. To address endogeneity, we introduce a new instrument: each country s number of UNESCO World Heritage Sites. Results suggest there is a robust positive (causal) relationship between tourism receipts and growth.

7 UNESCO World Heritage Sites: selection process UNESCO encourages the identification, protection, and preservation of cultural and natural heritage. To be eligible for sites, a country must be a signatory of the World Heritage Convention; as of 2006, 181 countries have signed. Countries nominate sites, and committee selects annually those to be added to the World Heritage List (WHL).

8 UNESCO WHL and tourism Testimony effect : WHL rewards sites the world recognizes as exceptional (e.g. Pyramids). Advertising i effect : WHL designation increases the visibility and accelerates the flow of tourists to a particular site. A3

9 Slide 7 A3 Example? ACHERIF, 10/13/2009

10 Validity of the Instrument t We discuss three potential sources of endogeneity of the suggested instrument: 1. Scope: Is there evident bias in the regional and historical distribution of World Heritage sites? 2. Political clout: Does political clout, with potential economic consequences, influence the WHL inscription process? 3. Cultural l vs. natural sites: Do natural sites affect economic performance through channels other than tourism?

11 Scope: global l coverage

12 Scope: historical i bias? Table 1. Regional and Historical Distribution of World Heritage Sites (2002) Region Cultural Natural 1/ Total Xth B.C.-XIVth A.D. XVth-XVIIth XVIIIth-XIXth XXth Africa (Early man, Islamic) (Zimbabwes) (Colonial) Asia (Buddhist, Hindu) (Ming, Mughal) (Qing) Middle East (Mesopotamia, Egypt, Islamic) (Ottoman) (North Africa) Europe (Greece, Rome, Middle Ages) 2/ (Renaissance) (Enlightenment, Industrial Rev.) Latin America (Aztec, Inca, Maya) (Spanish, Portuguese) (Independence) Total Source: UNESCO, / Includes "mixed" sites, i.e. those sites classified under both natural and cultural criteria. 2/ The Middle Ages account for 143 of the 219 European sites during this timeframe.

13 Political clout: UNESCO sites traded d for UN votes? Table 2. Correlation Between Total UNESCO World Heritage Sites and Average UN Voting Coincidence, Correlation coefficients for all countries (except G7) with: Barro & Lee (2005) Kegley & Hoock (1991) Thacker (1999) Canada France Germany Italy Japan UK USA G Correlation coefficients for non- OECD countries with: Canada France Germany Italy Japan UK USA G

14 Cultural l vs. natural sites Do natural sites capture natural resource abundance? Some natural sites are part of a recent conservationist i t trend. The instrument t could capture institutional variables.

15 Estimation Use instrumental variables technique to estimate standard growth models with a measure for tourism specialization (using WDI data). First stage: Tourism = UNESCO + Income + Education + Distance + Ɛ As well as instrumentation for institutions and trade openness. Second stage: Growth = Tourism + Income + Education + Distance + Ɛ

16 Benchmark regressions: first stage First Stage (1) (2) (3) (4) VARIABLES Tourism Tourism Tourism Trade Tourism Institution Unesco *** *** *** *** [2.851] [6.751] [6.863] [23.662] [7.066] [0.400] Income * ** 8.072** *** [0.551] [0.605] [3.725] [0.620] [0.130] Education [1.224] [1.206] [3.803] [1.301] [0.157] Distance *** [0.038] [0.042] [0.248] [0.035] [0.008] Kprice lnfrinstex_dk 1.421* *** [0.834] [5.786] engfrac_dk [3.842] [0.306] Constant 8.162*** ** *** * 9.613* * [1.028] [5.053] [6.304] [20.294] [5.206] [0.860] F test P value Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1

17 Benchmark regressions: second stage Second Stage (1) (2) VARIABLES Growth Growth (3) Growth (4) Growth Tourism 0.015*** 0.013** [0.005] [0.006] Income [0.069] Education 0.158** [0.066] Distance 0.013*** [0.004] Kprice 0.012** 0.015*** [0.006] [0.006] ** * [0.069] [0.143] 0.163** 0.205*** [0.071] [0.075] 0.015*** [0.004] [0.008] Trade [0.003] Institution Constant [0.067] [0.433] [0.203] [0.423] [0.609] Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (15% maximal IV size) Stock-Yogo weak ID test critical values (20% maximal IV size) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<

18 Robustness Use various versions of the WHL. Use exclusively the number of cultural sites, and subtract cultural sites by centuries of creation. Check results with another instrument: km of coastline times distance to equator (over-identification test).

19 Robustness (cont.) Use different definitions and data sources for the dependent variable. Re-estimation estimation using first differences ( advertising effect vs. testimony effect ). Test whether outliers drive the results. Augmented with controls for persistence effects (see Acemoglu et al. (2001), Tabellini (2007), and Guiso et al. (2008)).

20 Conclusion An increase of one standard deviation in tourism activity (8 percent of exports) leads to annualized additional growth of 0.5 percentage points. Can tourism make a miracle? The magnitude found makes it difficult. Tourism s impact on technology adoption, labor reallocation and learning-by-doing could explain why.

21 Using various list versions Table 3. Robustness using Various WHL (1) (2) (3) VARIABLES growth growth growth Tourism 0.013** 0.013* 0.015** [0.006] [0.007] [0.006] Income [0.069] [0.070] [0.070] Education 0.158** 0.158** 0.156** [0.066] [0.067] [0.066] Distance 0.013*** 0.013*** 0.013*** [0.004] 004] [0.004] 004] [0.004] 004] Constant [0.433] [0.435] [0.435] Cut-off year for instrument Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (15% maximal IV size) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1

22 Dropping natural sites Table 4. Robustness to Using Only Cultural Sites (1) (2) VARIABLES growth growth Tourism 0.013** 0.015*** [0.006] [0.005] Income [0.069] [0.070] Education 0.158** 0.156** [0.066] [0.067] Distance 0.013*** 0.013*** [0.004] [0.004] Constant [0.433] [0.437] Instrument coverage overall cultural only Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (15% maximal IV size) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1

23 Various century cut-offs Table 5. Robustness to Removing Various Centuries from the WHL (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES growth growth growth growth growth growth growth growth Tourism 0.015*** 0.015*** 0.016*** 0.017*** 0.017*** 0.018*** 0.019*** 0.022*** [0.005] [0.005] [0.005] [0.005] [0.005] [0.005] [0.005] [0.006] Income [0.070] [0.070] [0.070] [0.070] [0.070] [0.071] [0.071] [0.071] Education 0.156** 0.156** 0.154** 0.152** 0.151** 0.150** 0.148** 0.145** [0.067] [0.067] [0.066] [0.066] [0.066] [0.067] [0.067] [0.067] Distance 0.013*** 0.013*** 0.013*** 0.013*** 0.013*** 0.013*** 0.013*** 0.013*** [0.004] [0.004] [0.004] [0.004] [0.004] [0.004] [0.004] [0.004] Constant t [0.437] [0.437] [0.438] [0.439] [0.439] [0.442] [0.442] [0.445] Century Cut-off Point for Cultural Sites: All XX XVIII XV XIII X V V B.C. Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (15% maximal IV size) Stock-Yogo weak ID test critical values (20% maximal IV size) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1

24 Using coastline times distance Table 6. Robustness to Using Additional Instruments for Tourism (1) (2) (3) (4) VARIABLES growth growth growth growth Tourism 0.013** 0.015** 0.016** 0.015** [0.006] [0.006] [0.006] [0.006] Income [0.069] 069] [0.070] 070] [0.069] 069] [0.069] 069] Education 0.158** 0.155** 0.154** 0.155** [0.066] [0.066] [0.066] [0.066] Distance 0.013*** 0.013*** 0.013*** 0.013*** [0.004] [0.004] [0.004] [0.004] Constant [0.433] [0.434] [0.432] [0.429] Instrument coverage unesco unesco, coastline unesco, coastline, coastline interacted with distance unesco, coastline, coastline interacted with distance, and coastline squared Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (5% maximal IV size) Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (20% maximal IV size) Hansen J test (p value) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1

25 Using PPP GDP from PWT Table 7. Robustness to using Different Measures of GDP (1) (2) (3) (4) (5) GDP, PPP (constant 2005 Real GDP Chain per worker unit: I$ Real GDP per capita (Constant Prices: Real GDP per capita (Constant Prices: Chain series) unit: I$ in 2000 Real GDP Chain per equivalent adult unit: I$ international $) per worker in 2000 Constant Prices, Laspeyres) unit: I$ in 2000 Constant Prices, PWT per eq. adult in 2000 Describtion World Bank PWT 6.2 Constant Prices, PWT Constant Prices VARIABLES growth growth_rgdpwok_pwt growth_rgdpl_pwt growth_rgdpch_pwt growth_rgdpeqa_pwt Tourism 0013** *** 0021*** *** 0022*** [0.006] [0.007] [0.007] [0.007] [0.007] Income [0.069] Income_rgdpwok_pwt ** [0.070] Income_rgdpl_pwt [0.081] Income_rgdpch_pwt [0.081] Income_rgdpeqa_pwt [0.082] Education 0.158** * 0.149* 0.139* [0.066] [0.071] [0.079] [0.079] [0.079] Distance 0.013*** 0.015*** 0.012*** 0.012*** 0.012*** [0.004] 004] [0.004] 004] [0.004] 004] [0.004] 004] [0.004] 004] Constant [0.433] [0.511] [0.562] [0.561] [0.586] Kleibergen-Paap rk Wald F statistic Stock-Yogo weak ID test critical values (10% maximal IV size) Stock-Yogo weak ID test critical values (15% maximal IV size) Stock-Yogo weak ID test critical values (20% maximal IV size) Observations R-squared Robust standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1