To these days a number of empirical studies clearly show that tourism activity is strictly interconnected both with the level and the dynamics of regional GDP. As regional economy expands so does the tourism sector. Yet, if the linkages between tourism and the level of economic activity has been proved to be strong and robust, the causal relationship is still a puzzle (Pablo-Romero & Molina, 2013). Is it the flourishing of tourism activities to spur economic growth, or rather is it economic growth and development which allow the tourism sector to expand and to grow? This issue is still under investigation both empirically and theoretically. In fact if, for example, Balaguer and Cantavella-Jorda (2002) for Spain and Gunduz and Hatemi-J (2005) for Turkey show that tourism is actually a driver for economic growth, Narajan (2002) for Fiji and Oh (2005) for South Corea, find instead that it is economic growth which actually feeds the growth and the expansion of the tourism sector. The empirical evidence in many instances is even fuzzier since a clear causal direction is not easily identifiable and rather data point to a double causal relationship between tourism and economic growth (Dritsakis, 2004; Lee & Chien, 2008). Problems of measurement and identification of the variables to study add further difficulties to the investigations on the issue. In fact many studies have pointed out that these empirical results are very sensitive both to the model specification and to the employed econometric technique. Even if the real nature of the relationship and the degree of interconnection between tourism and level of aggregate economic activity is in general under scrutiny, very few can deny that tourism can play a crucial role in determining specifically the growth of underdeveloped regions, in particular of those regions which have a rich unexploited cultural heritage (Holzner, 2011; Rosentraub & Joo, 2009). Indeed, Cultural heritage, to a certain extent, can work as physical capital or natural resources in the production process and contribute significantly to the economy overall (Wagner 1997; Zhou, Yanagida, Chakravorty, & Leung 1997). The difficulty in identifying the nature of the interconnection between tourism demand and the main aggregate regional variables depends on the fact that there is a high number of channels through which this interaction may actually occur. Hence, identifying those channels and measuring the possible impact of a change in tourism expenditure on the economy is extremely important not only for theory purposes but also because this is the only way one can design optimal growth enhancing policies. Empirical investigations have proved that input-output analyses can serve very well in clearly mapping and highlighting the main channels through which tourism may impact the economy. In fact input-output analysis provides a useful description of the working of an economic system through the measurement of the exchange of resource flows among all sectors in the economy. In its simple formulation, each sector is considered to be a producer that supplies goods to all other remaining sectors, and, at same time, it is considered to be a consumer demanding goods from other sectors. However, this methodology can oversimplify matter and miss important linkages. For this reason more recently, many investigations have departed from the simple input-output framework and have started operating through richer models such as Computable General Equilibrium (CGE) models. The reason is that CGE models have the advantage of providing a richer picture of the interconnections between economic sectors taking into account income feedbacks, resource limitations and price adjustments. By departing from the recent efforts of part of the literature in providing more reliable and significant regional input-output tables, this paper seeks to combine the input-output approach with the agent-based approach. The main idea is that by combining agent-based models with the input-output analysis one can obtain more powerful explanatory frameworks. These frameworks can be a response, from one hand, to a greater need for flexibility and descriptive adaptability, and from another, to a need for computation. In fact, computational codes possess both formal requirements and adaptability and flexibility and, ultimately computability. The agent-based approach analyses the possible interactions in terms of communication and exchange among a large number of agents, and the interaction of them with the surrounding environment in which they are placed (Railsback & Grimm, 2012). The agents, modeled by choices of the researcher, are programmed to act in a way similar to entities as persons, institutions and firms, trying to reproduce their characteristics and their behaviour. (Kirman, 2010). The main task of the programmer-researcher is to define agents' features and capabilities, the actions that can be performed and the characteristics of the environment in which agents are placed, and, where appropriate, the effects of their actions on the environment itself (Gilbert & Terna, 2000). This paper has two objectives. The first is a general methodological objective. In the specific, the aim is to exploit the contribution of the agent-based modeling in order to make less stringent assumptions than the input-output analysis and to determine more realistic calculation of the economic impact. The second is an empirical objective. By applying this methodology the work aims to investigate the effects of a change in tourism demand on the economy of a specific Italian region, Campania. The final objective is to measure the impact on regional GDP and on the level of employment of a specific change in tourism demand. One expects this exercise to be particularly useful in providing useful policy recipes for promoting growth and development in a relatively underdeveloped region extremely rich in terms of cultural heritage. The empirical methodology is the following. In the absence of regional input-output tables, we first build regional input-output tables using various non-survey methods. These tables will then be used to model agents' attributes and behaviours in a fully integrated general equilibrium agent-based model. In the transition from input-output logic to agent-based logic, we use the data provided by the input-output table as parameters, we build agents with their attributes and behaviours, and finally, we model the relationships of monetary exchange between them. More specifically the work identifies one agent for each industrial sector (20 different industrial sectors and hence agents), one agent corresponding to the families, one to the regional government and state government, and finally, one corresponding to the foreign sector. Once agents have been identified, one can move on in modeling their attributes and behaviours, as well as the economic relations between them, using as parameters the statistical information available in the input-output table. This model will finally be tested by simulating an exogenous shock in the tourism demand.
The joining between input-output approach and agent-based modeling: measuring the impact of tourism in Campania
Luigi Guadalupi;Francesco Andreottola;
2015
Abstract
To these days a number of empirical studies clearly show that tourism activity is strictly interconnected both with the level and the dynamics of regional GDP. As regional economy expands so does the tourism sector. Yet, if the linkages between tourism and the level of economic activity has been proved to be strong and robust, the causal relationship is still a puzzle (Pablo-Romero & Molina, 2013). Is it the flourishing of tourism activities to spur economic growth, or rather is it economic growth and development which allow the tourism sector to expand and to grow? This issue is still under investigation both empirically and theoretically. In fact if, for example, Balaguer and Cantavella-Jorda (2002) for Spain and Gunduz and Hatemi-J (2005) for Turkey show that tourism is actually a driver for economic growth, Narajan (2002) for Fiji and Oh (2005) for South Corea, find instead that it is economic growth which actually feeds the growth and the expansion of the tourism sector. The empirical evidence in many instances is even fuzzier since a clear causal direction is not easily identifiable and rather data point to a double causal relationship between tourism and economic growth (Dritsakis, 2004; Lee & Chien, 2008). Problems of measurement and identification of the variables to study add further difficulties to the investigations on the issue. In fact many studies have pointed out that these empirical results are very sensitive both to the model specification and to the employed econometric technique. Even if the real nature of the relationship and the degree of interconnection between tourism and level of aggregate economic activity is in general under scrutiny, very few can deny that tourism can play a crucial role in determining specifically the growth of underdeveloped regions, in particular of those regions which have a rich unexploited cultural heritage (Holzner, 2011; Rosentraub & Joo, 2009). Indeed, Cultural heritage, to a certain extent, can work as physical capital or natural resources in the production process and contribute significantly to the economy overall (Wagner 1997; Zhou, Yanagida, Chakravorty, & Leung 1997). The difficulty in identifying the nature of the interconnection between tourism demand and the main aggregate regional variables depends on the fact that there is a high number of channels through which this interaction may actually occur. Hence, identifying those channels and measuring the possible impact of a change in tourism expenditure on the economy is extremely important not only for theory purposes but also because this is the only way one can design optimal growth enhancing policies. Empirical investigations have proved that input-output analyses can serve very well in clearly mapping and highlighting the main channels through which tourism may impact the economy. In fact input-output analysis provides a useful description of the working of an economic system through the measurement of the exchange of resource flows among all sectors in the economy. In its simple formulation, each sector is considered to be a producer that supplies goods to all other remaining sectors, and, at same time, it is considered to be a consumer demanding goods from other sectors. However, this methodology can oversimplify matter and miss important linkages. For this reason more recently, many investigations have departed from the simple input-output framework and have started operating through richer models such as Computable General Equilibrium (CGE) models. The reason is that CGE models have the advantage of providing a richer picture of the interconnections between economic sectors taking into account income feedbacks, resource limitations and price adjustments. By departing from the recent efforts of part of the literature in providing more reliable and significant regional input-output tables, this paper seeks to combine the input-output approach with the agent-based approach. The main idea is that by combining agent-based models with the input-output analysis one can obtain more powerful explanatory frameworks. These frameworks can be a response, from one hand, to a greater need for flexibility and descriptive adaptability, and from another, to a need for computation. In fact, computational codes possess both formal requirements and adaptability and flexibility and, ultimately computability. The agent-based approach analyses the possible interactions in terms of communication and exchange among a large number of agents, and the interaction of them with the surrounding environment in which they are placed (Railsback & Grimm, 2012). The agents, modeled by choices of the researcher, are programmed to act in a way similar to entities as persons, institutions and firms, trying to reproduce their characteristics and their behaviour. (Kirman, 2010). The main task of the programmer-researcher is to define agents' features and capabilities, the actions that can be performed and the characteristics of the environment in which agents are placed, and, where appropriate, the effects of their actions on the environment itself (Gilbert & Terna, 2000). This paper has two objectives. The first is a general methodological objective. In the specific, the aim is to exploit the contribution of the agent-based modeling in order to make less stringent assumptions than the input-output analysis and to determine more realistic calculation of the economic impact. The second is an empirical objective. By applying this methodology the work aims to investigate the effects of a change in tourism demand on the economy of a specific Italian region, Campania. The final objective is to measure the impact on regional GDP and on the level of employment of a specific change in tourism demand. One expects this exercise to be particularly useful in providing useful policy recipes for promoting growth and development in a relatively underdeveloped region extremely rich in terms of cultural heritage. The empirical methodology is the following. In the absence of regional input-output tables, we first build regional input-output tables using various non-survey methods. These tables will then be used to model agents' attributes and behaviours in a fully integrated general equilibrium agent-based model. In the transition from input-output logic to agent-based logic, we use the data provided by the input-output table as parameters, we build agents with their attributes and behaviours, and finally, we model the relationships of monetary exchange between them. More specifically the work identifies one agent for each industrial sector (20 different industrial sectors and hence agents), one agent corresponding to the families, one to the regional government and state government, and finally, one corresponding to the foreign sector. Once agents have been identified, one can move on in modeling their attributes and behaviours, as well as the economic relations between them, using as parameters the statistical information available in the input-output table. This model will finally be tested by simulating an exogenous shock in the tourism demand.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
