Exploring the tourism markets’ convergence hypothesis in Italy

Abstract

This article aims at investigating the tourism markets’ convergence hypothesis across Italy’s 20 major source markets. To reach our goal, we use monthly data of tourist arrivals and overnights over the period 2008–2018 and the time-varying factor model developed by Phillips and Sul (2007, 2009). Our findings suggest the absence of full (absolute) convergence, leading us to accept the hypothesis of club convergence. We show that the traditionally more important source markets have a tendency to persist, while Asian countries show heterogeneous behaviour. Furthermore, the relative decline in the contribution to total arrivals and overnights of several international source markets calls for a reconsideration of the promotional strategies to stimulate inbound tourism from these countries.

Keywords

convergence clubs Italy market segmentation tourism markets

JEL classification: L83, Z30, C23

Introduction

Tourism has become one of the largest and fastest growing economic sectors in the world due to its continued expansion and diversification (OECD, 2018). Particularly, the tourism sector offers an important contribution to the Italian economy and data regarding international tourism have recorded positive trends in recent years. Italy is the sixth country in the world for income from international tourism. It is the fifth country in the world for visitors and its tourism balance surplus is higher than the European average. In 2018, foreign travellers’ annual growth jumped to 6.5% and the increase in tourism expenditures reached 41.7 billions of Euro, equal to 40% of the exports of services (Bank of Italy, 2019).

The increasing importance of tourism in Italy has stimulated growing research on its role in affecting economic growth (i.e. Cortés-Jiménez, 2008; Giambona and Grassini, 2020; Paci and Marrocu, 2014). While from the theoretical and empirical points of views, there is a lack of consensus on whether tourism promotes economic activity or, conversely, economic activity leads to tourism growth, the tourism-led economic growth hypothesis seems to be the most evident for Italy (Antonakakis et al., 2015; Pablo-Romero and Molina, 2013).

Although the role of tourism in affecting the economic growth of Italy has been extensively analysed, the tourism markets’ convergence hypothesis has not yet been explored. This hypothesis discussed for the first time by Narayan (2006) has become an emerging topic in the recent tourism economics literature (see, e.g. Abbott et al., 2012; Kourtzidis et al., 2018; Lean and Smyth, 2008; Lin et al., 2019a; Mérida et al., 2016; Solarin, 2018).¹ Specifically, testing for tourism markets’ convergence allows to evaluate the performance of the tourism industry in a particular country as well as the success of policy strategies developed to boost the number of tourist arrivals. Such analysis may be also useful for segmenting global tourism markets and generates important implications for policymakers and tourism organizations to develop global marketing strategies.

Within this framework, this article aims at investigating the tourism markets’ convergence hypothesis across Italy’s 20 major source markets. To reach our goal, we use monthly data of tourist arrivals over the period 2008–2018 and the time-varying factor model developed by Phillips and Sul (2007, 2009) (PS henceforth). We repeat the analysis using also overnights, to better disentangle the evolution patterns of such source markets.

Our findings suggest the absence of full (absolute) convergence, leading us to accept the hypothesis of club convergence. The results show the significance of traditionally important source markets (Western European countries and the United States) and the growing importance of some Asian source markets (China and Korea). In addition, we show the transitional divergence of other source markets that call for a reconsideration of the strategies to increase tourist arrivals and overnights from these countries.

Data and methodology

To study tourism markets’ convergence, we use the number of tourist arrivals and overnights of Italy’s 20 major source markets for the period 2008m1–2018m12. Data are taken from the Italian National Institute of Statistics.² Table 1 presents the countries included in the analysis together with the average values of our outcome variables throughout the whole period. Germany is undoubtedly the major source market of Italy, followed by the United States. The European-core countries are of key importance as well, since they contribute to a great extent to the international tourism inflow. Conversely, other source markets appear lagging behind either in terms of arrivals or in terms of overnights (i.e. Brazil, Canada and Korea).

Table 1.

Descriptive statistics.

Country	Arrivals	Overnights	Length of stay
Australia (AUS)	839,422	2,190,760	2.6
Austria (AUT)	2,198,689	8,599,643	3.9
Belgium (BEL)	1,120,583	4,674,652	4.2
Brazil (BRA)	738,179	1,811,659	2.5
Canada (CAN)	723,953	1,977,112	2.7
China (CHN)	1,979,889	3,167,869	1.6
Czech Republic (CZE)	662,785	3,270,787	4.9
Germany (DEU)	10,447,577	52,154,374	5.0
Denmark (DNK)	619,967	3,282,387	5.3
Spain (ESP)	1,846,499	4,923,386	2.7
France (FRA)	3,922,535	11,970,936	3.1
United Kingdom (GBR)	3,127,320	11,881,800	3.8
Japan (JPN)	1,232,914	2,432,006	2.0
Korea (KOR)	566,682	1,020,668	1.8
Netherlands (NLD)	1,962,221	10,831,098	5.5
Poland (POL)	1,076,945	4,271,004	4.0
Romania (ROU)	536,046	2,065,976	3.9
Russia (RUS)	1,398,423	5,014,559	3.6
Sweden (SWE)	626,536	2,320,899	3.7
United States (USA)	4,555,292	11,676,847	2.6

Note: Values indicate annual averages over the period 2008–2018. Length of stay is equal to the ratio Overnights/Arrivals.

As for the analysis, we follow the methodology developed by Phillips and Sul (2007, 2009). They propose to modify the conventional panel data decomposition of the variable of interest (in our case, international tourist arrivals or overnights) in the following way:

y_{i t} = b_{i t} μ_{t}

where b_it is the systematic idiosyncratic element that is allowed to evolve over time and μ_t is the common factor.³ To test if different tourism source markets converge, we define the following relative transition component for each outcome variable:

h_{i t} = \frac{y_{i t}}{N^{- 1} \sum_{i = 1}^{N} y_{i t}} = \frac{b_{i t}}{N^{- 1} \sum_{i = 1}^{N} b_{i t}}

that can be directly computed from the data. In such a way, it is possible to remove the common steady-state trend μ_t, tracing an individual trajectory for each economy i in relation to the panel average. The factor μ_t may represent the aggregated common behaviour of y_it, but it could also be any common variable affecting the individual behaviour.

The appealing feature of the PS approach is that such a factorization allows the separation of unit-specific transitional factors from common factors that allow to reveal the long-run trend for the underlying time series. In other words, by using this approach, we are able to control for the common factors that affect the behaviour of tourist arrivals (and overnights) from different countries of origin, and, therefore, we can focus on their idiosyncratic behaviour.

In presence of convergence, the coefficient h_it should converge towards unity: $h_{i t} \to 1$ , for all i, as $t \to \infty$ . At the same time, the cross-sectional variation H_it (computed as the quadratic distance measure for the panel from the common limit) should converge to zero:

H_{t} = N^{- 1} \sum_{i = 1}^{N} {(h_{i t} - 1)}^{2} \to 0 as t \to \infty

To formally test the presence of convergence, PS suggests to estimate the following equation by ordinary least squares methodology:

log \frac{H_{1}}{H_{t}} - 2 log (log t) = a + β log t + u_{t}

for t = [rT], [rT] + 1,…, T, where β represents the speed of convergence parameter of b_it; r assumes a positive value to discard the first block of observation from the estimation and [rT] is the integer part of rT. The null hypothesis of convergence is tested through a one-sided t-test robust to heteroskedasticity and autocorrelation and specifically, it is rejected at the 5% level if $t_{\hat{β}} < - 1.65$ .

This procedure, generally called ‘log t-test’, has power against cases of club convergence and it is one of the most important features of the PS approach. Specifically, if the log t-test is rejected for the whole sample, the authors suggest to repeat the test procedures according to a clustering mechanism concerning of four steps (for details, see Phillips and Sul, 2007). To test the robustness of our results, we also apply the club merging algorithm initially developed in Phillips and Sul (2009) and a different club merging algorithm recently developed by von Lyncker and Thoennessen (2017).⁴

The PS framework has additional advantages over other existing methodologies. First, if the full panel of economies analysed does not converge to a common steady state, the methodology allows to investigate for the presence of groups of economies (in our case, source markets) that converge to different equilibria, and at the same time, it permits individual economies to diverge. Second, under such (possible) transitional paths heterogeneity, standard unit root and cointegration tests tend to be not suitable for examining convergence and, as demonstrated in Phillips and Sul (2007), even if two series are not cointegrated, it is still possible to find convergence between them. Finally, unlike other approaches in which economies are grouped a priori, the methodology enables the endogenous determination of convergence clubs if the log t-test is rejected for the whole sample.

Results

When we apply the log t-test to monthly arrivals of Italy’s 20 major international source markets over the period 2008–2018, the hypothesis of absolute convergence is rejected at the 1% significance level ( $t_{\hat{β}} < - 273.96$ ).⁵ Therefore, we follow the recursive algorithm proposed by PS to verify the presence of convergence clubs. Table 2 (panel A) shows that the application of the algorithm yields two convergence clubs for which $t_{\hat{β}} > - 1.65$ , thus indicating the presence of significant (though weak) convergence within each club. Results do not change when we apply the two club merging algorithms earlier discussed.

Table 2.

Convergence club classification.

Panel A – Arrivals
	Club	Members	n° Countries	β	SE	t-stat
	1	USA, FRA, GBR, CHN, DEU, ESP, AUT, POL, KOR	9	0.227	0.007	31.763
	2	NLD, RUS, BRA, JPN, BEL, AUS, ROU, CAN, CZE, SWE, DNK	11	0.227	0.011	20.571
Panel B – Overnights
	Club	Members	n° Countries	β	SE	t-stat
Baseline PS algorithm	1	USA, FRA, GBR, CHN, KOR	5	0.621	0.009	67.898
	2	ESP, RUS, POL, AUT, NLD, BEL, BRA	7	0.059	0.004	14.348
	3	CZE, ROU, AUS, JPN, SWE, CAN, DNK	7	0.283	0.011	25.626
	Club	Members	n° Countries	β	SE	t-stat
Club merging algorithms	1	USA, FRA, GBR, CHN, KOR, ESP, RUS, POL, AUT, NLD, BEL, BRA	12	0.323	0.009	35.783
Club merging algorithms	2	CZE, ROU, AUS, JPN, SWE, CAN, DNK	7	0.283	0.011	25.626

Note: USA: the United States; FRA: France; GBR: the United Kingdom; CHN: China; DEU: Germany; ESP: Spain; AUT: Austria; POL: Poland; KOR: Korea; NLD: the Netherlands; RUS: Russia; BRA: Brazil; JPN: Japan; BEL: Belgium; AUS: Australia; ROU: Romania; CAN: Canada; CZE: Czech Republic; SWE: Sweden; DNK: Denmark. DEU is diverging country in the case of overnights.

We find that the first club identified is prevalently composed of traditionally important source markets (such as Germany, the United States, France and the United Kingdom), additional European countries for which proximity encourages people to visit Italy (Austria, Spain and Poland) and emerging Asian markets (China and Korea). On the other hand, the second club pertains to other European countries (the Northern–Eastern ones) and most of the other major international source markets (Australia, Brazil, Canada, Japan and Russia).

Interesting insights emerge when looking at Figure 1 that describes the relative transition paths within each group as well as the average relative transition paths of both clubs. Traditional source markets tend to maintain their importance over time. Among the non-European advanced countries, the US share has fluctuated over time reflecting the dynamics of exchange rates. Conversely, Asian markets in club 1 (China and Korea) show increasing importance in the relative share of tourist arrivals. This should be not surprising since Italy has become one of the main destinations for Korean visitors (World Tourism Organization, 2018). The growing income levels in these countries, the reduction of transport costs and the new types of tourist supply through accommodation rental platforms (i.e. Airbnb, Booking) can explain the continuous increase in tourism arrivals from Asia.

Figure 1.

Relative transition paths within and across clubs – arrivals.

On the contrary, almost all source markets in club 2 show a sort of decline in the relative importance of tourist arrivals (exceptions are Romania and Brazil). In particular, the behaviour of three source markets is noteworthy. Brazil still remains marginal for Italy with respect to other emerging and developing markets included in our sample, though it is gaining relevance. On the other hand, the importance of the Japanese market has declined in the latest years recording a drop in arrivals (−35% in the period 2012–2018), despite the fact that Italy is one of the favourite destinations of Japanese tourists. This appears a direct consequence of the persistent stagnation suffered by the Japanese economy during the past decade. Less clear, instead, is the behaviour of the Russian market. Tourism flows from Russia reached the peak in the 2012–2013 period, recording a significant drop in the following years as a consequence of the recession and the rouble devaluation.

As a result, the third panel of Figure 1 shows the divergent behaviour between source markets gathered in club 1 with respect to those belonging to club 2. One of the reasons of such divergence could be the way in which the countries of origin were affected by the financial crisis started in 2008, but further analysis would be needed to confirm such a conclusion. It seems that the countries belonging to club 1 recovered much better than those of club 2 and this may have affected tourist decisions. In the period 2012–2018, the average growth rate in real gross domestic product per capita for countries of club 1 was about 15%, while it was 9% for countries of club 2.⁶

To better disentangle the contribution of each source market, we also performed the analysis using monthly data of overnights. Also, in this case, the hypothesis of absolute convergence is rejected at the 1% significance level $(t_{\hat{β}} = - 137.03)$ . Table 2 (panel B) presents the results obtained by applying the baseline algorithm proposed by PS and those of the two club merging algorithms earlier discussed. The results obtained in the final classification mostly confirm the club composition already found for the arrivals. Notable exceptions are Belgium, Brazil, Russia and The Netherlands, that in this case are clustered in the first club instead of in the second one. In addition, the transitional convergence of China and Korea towards the highest source markets in the first club is complemented by that of Brazil, and the divergence between the two clubs is even more marked in this case (Figure 2).

Figure 2.

Relative transition paths within and across clubs – overnights.

Conclusions and policy implications

In this article, we have empirically examined the presence of the tourism markets’ convergence hypothesis in the case of Italy. The results suggest the absence of full (absolute) convergence in the case of both arrivals and overnights, in favour of the presence of convergence clubs.

The market segmentation we found analysing data from the last decade (2008–2018) can be only partly explained by geographical distance; instead, it seems related to the resilience of countries of origin to the financial crisis of 2008. Therefore, strategies devoted to increase inbound tourism from the nearest countries in geographical terms may be not fully satisfactory to increase the market share of Italy in the tourism sector.

From this analysis, several key findings emerged. First, the traditionally more important source markets have had a tendency to persist even after the Great Recession and, therefore, marketing strategies planned for these countries seem to be appropriate.

Second, Asian countries clearly represent a source of further growth for the Italian tourism sector but customized policies and strategies for each country are needed (i.e. the transitional convergence of Chinese and Korean source markets is counterbalanced by the transitional divergence of the Japanese market, which still reflects the persistent stagnation suffered by the Japanese economy during the past decade). A strengthening of the ability to intercept the expansion of the demand for tourist services from these markets that are growing rapidly but that are geographically and culturally more distant appears relevant for the Italian economy.

Third, the relative decline in the contribution to total arrivals and overnights of several source markets calls for a reconsideration of the promotional strategies outlined for these countries that appear to be no longer effective.

Particularly, since the divergence in tourist arrivals and overnights may be a function of the resilience to the slowdown caused by the Great Recession in the countries of origin, a strategy devoted to make tourist packages more accessible to mass tourism could be successful to regain market shares. At the same time, the diversification of the national offer through the promotion of interesting places with high potential (i.e. small villages, parks, mountain and protected areas, and rural areas) but still not fully recognized as international tourist destinations could be a viable way to stimulate inbound tourism from these countries.

On this point, further analysis would be needed also in the light of the severe cost imposed by the COVID-19 pandemic to the World economy and in particular to the tourism sector. The recent gains in tourist flows coming from the Asian markets as well as the large incoming flows from geographically more distant countries (i.e. the United States) could be compromised due to the likely increase in transport costs and because of the major blow that the leading accommodation rental platforms are facing.

Supplemental material

Supplemental Material, Appendix - Exploring the tourism markets’ convergence hypothesis in Italy

Supplemental Material, Appendix for Exploring the tourism markets’ convergence hypothesis in Italy by Pietro Pizzuto and Caterina Sciortino in Tourism Economics

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Pietro Pizzuto

Supplemental material

Supplemental material for this article is available online.

Notes

Author biographies

Pietro Pizzuto received his PhD in economics from the University of Palermo in 2017. His main research interests are in the area of regional convergence and resilience processes with particular attention to the nature of territorial imbalances. His research focuses also on the relationship between economic growth and tourism. He has published in several scientific journals including Economics Letters, Papers in Regional Science, Growth and Change and Journal of Economic Studies.

Caterina Sciortino received her degree in economics and finance from the University of Palermo in 2018. She is currently a PhD student (second year in progress) at the Department of Economics, Business and Statistics. Her research objectives are in the area of the Tourism Sciences, with particular attention to the economic impact of cruise tourism, the importance of tourist mobility factors, the survey methodology connected with tourism studies and the impact of behavioural economics in tourism scenario with particular focus on the decision process in travel choices.

References

Abbott

De Vita

Altinay

(2012) Revisiting the convergence hypothesis for tourism markets: evidence from Turkey using the pairwise approach. Tourism Management 33(3): 537–544.

Antonakakis

Dragouni

Filis

(2015) How strong is the linkage between tourism and economic growth in Europe? Economic Modelling 44: 142–155.

Bank of Italy (2019) Indagine sul turismo internazionale 2018. Available at: https://www.bancaditalia.it/pubblicazioni/indagine-turismo-internazionale/index.html (accessed on 4 March 2020).

Cortés-Jiménez

(2008) Which type of tourism matters to the regional economic growth? The cases of Spain and Italy. International Journal of Tourism Research 10(2): 127–139.

Giambona

Grassini

(2020) Tourism attractiveness in Italy: regional empirical evidence using a pairwise comparisons modelling approach. International Journal of Tourism Research 22(1): 26–41.

Kourtzidis

Tzeremes

, et al. (2018) Integration of tourism markets in Australia: an international visitor arrivals’ convergence assessment. Tourism Economics 24(7): 901–907.

Lean

Smyth

(2008) Are Malaysia’s tourism markets converging? Evidence from univariate and panel unit root tests with structural breaks. Tourism Economics 14(1): 97–112.

Lin

You

Lau

, et al. (2019a) Segmenting global tourism markets: a panel club convergence approach. Annals of Tourism Research 75: 165–185.

Lin

Yang

(2019b) Where can tourism-led growth and economy-driven tourism growth occur? Journal of Travel Research 58(5): 760–773.

10.

Mérida

Carmona

Congregado

, et al. (2016) Exploring the regional distribution of tourism and the extent to which there is convergence. Tourism Management 57: 225–233.

11.

Narayan

(2006) Are Australia’s tourism markets converging? Applied Economics 38(10): 1153–1162.

12.

OECD (2018) OECD Tourism Trends and Policies 2018. Paris: OECD Publishing.

13.

Ozcan

Erdogan

(2017) Are Turkey’s tourism markets converging? Evidence from the two-step LM and three-step RALS-LM unit root tests. Current Issues in Tourism 20(4): 425–442.

14.

Pablo-Romero

MDP

Molina

(2013) Tourism and economic growth: a review of empirical literature. Tourism Management Perspectives 8: 28–41.

15.

Paci

Marrocu

(2014) Tourism and regional growth in Europe. Papers in Regional Science 93: S25–S50.

16.

Phillips

Sul

(2007) Transition modeling and econometric convergence tests. Econometrica 75(6): 1771–1855.

17.

Phillips

Sul

(2009) Economic transition and growth. Journal of Applied Econometrics 24(7): 1153–1185.

18.

Sichera

Pizzuto

(2019) ConvergenceClubs: a package for performing the Phillips and Sul’s club convergence clustering procedure. The R Journal. DOI: 10.32614/RJ-2019-021.

19.

Solarin

(2018) Convergence hypothesis in tourism markets and activities in Taiwan. Tourism Economics 24(8): 1037–1044.

20.

von Lyncker

Thoennessen

(2017) Regional club convergence in the EU: evidence from a panel data analysis. Empirical Economics 52(2): 525–553.

21.

World Tourism Organization (2018) European Union Tourism Trends, UNWTO, Madrid. DOI: 10.18111/978928441947.0

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.12 MB