Tourism and local growth in Italy

ABSTRACT

Economic theory and the empirical literature are not conclusive on the relationship between tourism and economic growth. In this paper we estimate the impact of foreign tourists’ spending on value added per capita growth in the Italian provinces, using various econometric strategies. The overall results show that the effect is positive and statistically significant, but modest in economic terms. The impact is larger for the less developed provinces, and null for those that showed the highest tourist revenues per inhabitant at the beginning of the period, suggesting that congestion phenomena may occur.

KEYWORD:

• tourism
• economic growth
• tourism expenditure
• Beach Disease
• congestion
• instrumental variables

JEL:

• L83
• R0
• R10
• R11

ACKNOWLEDGEMENTS

The authors thank the participants at the Bank of Italy’s seminars, and at the 60th Annual Conference of the Italian Economic Association (SIE) held at the University of Palermo (October 2019) for their useful comments. They are grateful to Cristina Fabrizi for having shared the provincial housing price index with them, and to Monica Andini, Andrea Petrella, Paolo Sestito, Roberto Torrini and two anonymous referees for their valuable suggestions. The views and opinions expressed in this paper are those of the authors only and do not represent in any way those of the Bank of Italy.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Notes

1. We do not examine the potential relevance of spatial spillover because it is beyond the scope of our paper. We nevertheless tested for spatial correlation in the error term using an inverse distance spatial matrix (among provincial capitals). The Moran I and the LM tests, implemented by the Stata command spatdiag (Pisati, 2001), do not reject the null hypothesis of no spatial correlation at the 10% level (p > 0.10). The result is not sensitive to the boundary chosen for the maximum distance.

2. The value of 1.3 is an empirical rule of thumb motivated by a review of randomized experiments (Oster, 2019, p. 189).

3. The survey is carried out at the border (ports, airports, highways, etc.). Foreign tourists are asked the amount spent in the country visited, together with other information (the reason of travelling, total overnight stays, type of accommodations, etc.). If travellers visited more than one province, the total expenditure is imputed to each one in proportion of the numbers of overnight stays in each province (Bank of Italy, 2012, 2017a, 2017b).

4. The figures are not shown but are available from the authors upon request.

5. For the coefficients of the additional covariates, see Table A1 in Appendix A in the supplemental data online.

6. Three variables included in the covariates are measured as interpolations between 1991 and 2001. The results of column 4 are confirmed if we use their 1991 values.

7. The total number of nights that tourists spent in tourist accommodation.

8. Bronzini et al. (2020) suggest that the expansion of tourism in Rome that followed the Great Jubilee 2000 did not bring an increase in per capita value added.

9. Data between 2000 and 2005 are from Il Consulente Immobiliare (Muzzicato et al., 2008), while changes in the following years are estimated on data from the Osservatorio del mercato immobiliare.

10. In order to estimate housing supply elasticities for each Italian province, Accetturo et al. (2019) regress the growth rates of the housing stock on growth rates of house prices and some controls, allowing the coefficient on house prices (i.e., housing supply elasticities) to change across provinces.

11. The Hansen test is actually checking for the consistency of the result across different identification assumptions (different instruments). Although passing the Hansen test is crucial for interpreting the estimate as the effect of interest, it is not a test of validity of the moment restrictions (Parente & Santos Silva, 2012).

12. A not-too-large p-value is also reassuring of the fact that we do not have a problem of too many weak instruments, because in that case p-values tend to be extremely large (Roodman, 2009).

13. In 1997 foreign tourists’ average per inhabitant expenditure was €391, with value added per capita of €16,299.