How are Investments Allocated in a Publicly Owned Port System? Political Factors versus Economic Criteria

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=cres20 Regional Studies ISSN: 0034-3404 (Print) 1360-0591 (Online) Journal homepage: https://www.tandfonline.com/loi/cres20 How are Investments Allocated in a Publicly Owned Port System? Political Factors versus Economic Criteria JOSÉ I. Castillo-Manzano & Xavier Fageda To cite this article: JOSÉ I. Castillo-Manzano & Xavier Fageda (2014) How are Investments Allocated in a Publicly Owned Port System? Political Factors versus Economic Criteria, Regional Studies, 48:7, 1279-1294, DOI: 10.1080/00343404.2012.699192 To link to this article: https://doi.org/10.1080/00343404.2012.699192 Published online: 24 Jul 2012. Submit your article to this journal Article views: 368 View related articles View Crossmark data Citing articles: 2 View citing articles How are Investments Allocated in a Publicly Owned Port System? Political Factors versus Economic Criteria JOSÉ I. CASTILLO-MANZANO*and XAVIER FAGEDA† *Facultad de Ciencias Económicas y Empresariales, University of Seville, Av. Ramón y Cajal, 1, E-41018 Seville, Spain. Email: [email protected] †Department of Economic Policy, University of Barcelona, Av. Diagonal, 690, E-08034 Barcelona, Spain. Email: [email protected] (Received April 2011: in revised form May 2012) CASTILLO-MANZANO J. I. and FAGEDA X. How are investments allocated in a publicly owned port system? Political factors versus economic criteria, Regional Studies. This paper estimates the determinants of investments in port infrastructure in the Spanish regions made by the central administration. It is found that the use of the infrastructure in relation to capacity has some influence on the amount of investment received by a region. Furthermore, specialization in containers is also relevant for receiving more resources. However, it is found that when greater political decentralization does not go hand in hand with greater financial decentralization, the importance of tactical political aspects increases. Overall, efficiency plays a more relevant role in the regional allocation of investments in ports than in other transport infrastructures. Ports Investment Efficiency Solidarity Political factors Container transhipment CASTILLO-MANZANO J. I. and FAGEDA X. 公有港口系统如何配置投资？政治因素相对于经济准则， 区域研究 。本文评 价由西班牙中央政府决策的区域港口基础建设投资之决定因素。研究发现，基础建设的使用之于港口吞吐量，对于 一个区域所接受的投资数量具有某种程度的影响。此外，仓储的专殊化亦关乎获得更多的资源。但研究也发现，当 更多的政治分权无法配合更进一步的财政分权时，政治手段方面的重要性便会增加。总体来说，相较于其他交通基 础建设而言，效率在港口投资的区域配置方面有着更为关键的角色。 港口投资效率团结政治因素仓储转运 CASTILLO-MANZANO J. I. et FAGEDA X. Comment les investissements sont-ils alloués dans un système portuaire à caractère public? Les facteurs politiques par rapport aux critères économiques, Regional Studies. Cet article cherche à évaluer les déterminants des investissments dans l’infrastructure portuaire aux régions espagnoles réalisés par l’administration centrale. Il s’avère que l’utilisation de l’infrastructure en matière de capacité influe jusqu’à un certain point sur le montant de l’investissement qu’une région a attiré. Qui plus est, la spécialisation en conteneurs se révèle important pour attirer des ressources. Cependant, il s’avère que l’importance des aspects politiques tactiques augmente au moment où la décentralisation politique accrue ne va pas de pair avec la décentralisation financière. Globalement, l’efficacité joue un rôle plus important dans l’octroi régional des aides à l’investissement dans les ports plutôt que dans d’autres moyens de transport. Ports Investissement Efficacité Solidarité Facteurs politiques Transbordement de conteneurs CASTILLO-MANZANO J. I. und FAGEDA X. Wie werden Investitionen in einem öffentlichen Hafensystem verteilt? Politische Faktoren im Vergleich zu wirtschaftlichen Kriterien, Regional Studies. In diesem Beitrag werden die Determinanten für die Investitionen der Zentralregierung in die Hafen-Infrastruktur von spanischen Regionen untersucht. Wir stellen fest, dass die Auslastung der Infrastruktur im Vergleich zur Kapazität einen gewissen Einfluss auf den Umfang der von einer Region erhaltenen Investitionen ausübt. Darüber hinaus ist eine Spezialisierung auf Container ebenfalls für den Erhalt von zusätzlichen Ressourcen relevant. Allerdings stellen wir fest, dass die Bedeutung von taktischen politischen Aspekten zunimmt, wenn eine stärkere politische Dezentralisierung nicht mit einer stärkeren finanziellen Dezentralisierung einhergeht. Insgesamt spielt die Effizienz für die regionale Verteilung von Investitionen bei Häfen eine größere Rolle als bei anderen Verkehrsinfrastrukturen. Häfen Investitionen Effizienz Solidarität Politische Faktoren Containerumschlag Regional Studies, 2014 Vol. 48, No. 7, 1279–1294, http://dx.doi.org/10.1080/00343404.2012.699192 © 2012 Regional Studies Association http://www.regionalstudies.org CASTILLO-MANZANO J. I. y FAGEDA X. ¿Cómo se distribuyen las inversiones en un sistema portuario público?: Factores políticos frente a criterios económicos, Regional Studies. En este artículo se estiman los factores que determinan la distribución de las inversiones del gobierno central en el sistema portuario español. Se concluye que en este reparto se prima a aquellos puertos que más tráfico gestionan en relación a sus activos y a los que presentan una mayor especialización en el tráfico de contenedores. Sin embargo, los criterios políticos ganan importancia cuando la descentralización política no va acompañada de una descentralización financiera. En resumen, la eficiencia juega un papel más relevante en la distribución de las inversiones portuarias de lo que influye en el reparto de otras infraestructuras. Puertos Inversión Eficiencia Solidaridad Factores políticos Transbordo de contenedores JEL classifications: H54, L91, R42 INTRODUCTION Investments in transportation infrastructure can have a notable impact on regional economic growth. This is especially true when these investments are directed at territories, where citizens and firms make a greater use of them, whether to meet some unmet demand, to alleviate congestion problems or to correct bottlenecks in a given network. A number of empirical studies (for example, ALBALATE et al., 2012; CADOT et al., 2006; CASTELLS and SOLÉ-OLLÉ, 2005; KEMMERLING and STEPHAN, 2002, 2010; SOLÉ-OLLÉ, 2010, 2012) have analysed the determinants for central governments’regional allocations of investments in transportation infrastructure in a variety of European countries. Although there is a consensus regarding the economic importance of transport infrastructures, a common result in studies such as these is that the importance of considerations of efficiency or performance is modest for the decision about which territories are allocated a larger or a smaller proportion of the available resources for investment in infrastructure. As a result, issues of political tactics, pressure from certain interest groups or a preference for satisfying redistribution criteria play a role that is even more important than that played by the concern for greater economic efficiency. In the case of Spain, CASTELLS and SOLÉ-OLLÉ (2005) estimated an equation for the determinants of investment in general transportation infrastructure for a sample of regions in the country during the period 1987–1996. This study finds evidence of the existence of political factors that divert government attention to the efficiency–equity dilemma in investment decisions. Among these political factors, statistical significance is found for aspects such as the support given by regional political parties to the party that is governing in a minority in the central administration, the number of votes that the party in government needs in the region to gain a representative in parliament, and the fact that the same political party is in power in both the central and regional governments. 1 CADOT et al. (2006) simultaneously estimated an aggregated production function and an equation for the determinants in investment in infrastructure for a sample of regions in France during the period 1985– 1992. They concluded that the regions where the investment’s marginal productivity is higher do not necessarily receive greater investment, whilst the influence of certain interest groups (measured by the number of large companies based in the region) or political factors, such as the difference in votes between the two main political parties, do impact greatly on this decision. Meanwhile, KEMMERLING and STEPHAN (2002) also simultaneously estimated production and investment allocation functions for a sample of German towns and cities for 1980, 1986 and 1988. They found that redistribution considerations have a greater importance for decisions on infrastructure than any related to efficiency or performance, and those political factors, such as the fact that the same party in is power in local and regional governments, also play a significant role. KEMMERLING and STEPHAN (2010) emphasized the importance of country-specific political institutions in order to explain the regional distribution of investment. These studies use an aggregated focus, where the considerations of efficiency and solidarity are mainly related through the relationship between investments and levels of activity in the region. In particular, efficiency is generally measured as the ratio between gross domestic product (GDP) in relation to the capital stock and solidarity is measured through the GDP per capita. This implies that the use of the infrastructure, which should be measured by the levels of trafficit bears, is habitually included indirectly. Investments in the various modes of transportation are taken into account, but the analysis focuses especially on network infrastructure (road and rail) given the size of the investment made in both these modes. Note that the transportation literature usually considers efficiency in infrastructure transportation by relating traffic intensity to existent capacity. There are some specific studies for airports relative to the cases of Spain (BEL and FAGEDA, 2009) and the United States (BILOTKACH, 2010). As they examine a specific mode of transportation, their analyses of efficiency can include variables directly related to traffic. These studies find that tactical political factors seem to 1280 José I. Castillo-Manzano and Xavier Fageda be more important than economic factors. As far as the authors are aware, there is no study that has analysed the determinants for investment in ports. The main contribution of this article is the empirical analysis of the relative role that economic and political considerations play in investments in port infrastructure. 2 It is not clear whether in the sphere of ports political considerations are as important as for other modes of transportation. In fact, it might be anticipated a priori that performance could play a more important role. Only a limited number of regions have ports in most countries. This means that they are less attractive for policies such as territorial solidarity than other more omnipresent types of transport infrastructure such as airports and roads. They have, however, traditionally played a major role as regional development policy instruments. Regardless of the ownership structure, the management of many major world ports has been transferred from national or state governments to local governments, or even to the level of an individual entity. In this way, it has become more common for port authorities to have a high degree of financial independence for deciding on investments and the way that these are financed, especially over the last two decades. Most of empirical studies consider that the higher degree of autonomy in the management of ports has brought about gains in efficiency and performance. The higher degree of autonomy may be due to decentralization (BARROS, 2003; GONZÁLEZ and TRUJILLO, 2008), a higher involvement of private capital (CULLINANE et al., 2002; TONGZON and HENG, 2005) or both these reasons (ESTACHE et al., 2002). This could be taken as an indication that efficiency and port performance considerations have played a relevant role in the allocation of investments. Finally, port activity mainly focuses on the movement of freight, as a result of which investment in this type of infrastructure is less productive for securing votes compared with other infrastructure, such as roads, airports and railways where there are great movements of people (JACOBS, 2007). Furthermore, it should be noted that many port facilities have been set further away from city centres in recent decades to avoid congestion (BICHOU and GREY, 2005). This process has taken place in many old port cities that have had to renew facilities. Hence, the citizen-voter is less aware of port activities. This study focuses on an analysis of the determinants of investment in ports in Spain. The institutional framework is characterized by financial centralization and growing political decentralization within a port devolution process that is similar to that found in other Mediterranean countries. There are also financial instruments that redistribute resources among ports in the country. The purpose of this study is to test whether the existence of a financially centralized and politically decentralized institutional framework might also lead to other considerations apart from port performance playing an important role in the regional allocation of resources, as the previous literature states is the case for other transportation infrastructure. In particular, the authors want to test whether the central government favours more efficient ports (that is, those ports where the use of the infrastructure is more intensive), or whether the regional allocation of investment in ports is guided by other economic criteria, like inter-regional solidarity. Furthermore, the authors want to test whether some variables related to political tactics or the pressure of interest groups explain why a region receives more resources for its ports. In this regard, the central research question aims to examine whether economic criteria or political factors (or both) play a major role in explaining the regional allocation of investments in port facilities. In order to address this, multivariate regressions are performed to account of the quantitative impact of all the potential relevant factors that may determine public investments in ports. 3 In particular, the following questions are addressed: .Does the use of the infrastructure have any influence on future investments? .Is specialization by the Spanish port system in certain specific emerging traffic, such as container transport, being favoured? .Do mechanisms, such as the inter-port compensation fund, really allow funds to be redistributed from large to small ports? .Is investment being favoured by the inclusion of regional governments in the management of ports of the same political colour as the central government? .Has the relative weight in the regional economy of manufacturing and construction firms influence on the allocation of investments? The paper is organized as follows. The second section sets out the main features of the Spanish port system and focuses on the various factors that impact on the way investment is shared out among the country’s ports. The third section describes the empirical model used, including the origin of the data, and justifies the use of the variables. The fourth section explains the econometric issues that need to be considered in the regression analysis. The fifth section presents the results of the estimates. Finally, the sixth section discusses the findings and presents the conclusions that are drawn. INVESTMENT IN THE SPANISH PORT SYSTEM Spain has one of the most studied port devolution processes of the last decade (CASTILLO-MANZANO and ASENCIO-FLORES, 2012; CASTILLO-MANZANO et al., 2008; GONZÁLEZ and TRUJILLO, 2008). 4 This process has developed with four changes in legislation from How are Investments Allocated in a Publicly Owned Port System? 1281 the beginning of the 1990s to the present day, specifically Laws 27/1992, 62/1997, 48/2003 and 33/2010. These reforms broke with the traditional public model of ports in the hands of the central government which had been common to Mediterranean port systems in the preceding decades (SUYKENS and VAN DE VOORDE, 1998). Public ownership and centralized decision-making for investment were continued under the new model. However, privately owned terminals with a high degree of independence from the port authorities were permitted on port premises. Also, broadly speaking, a kind of decentralization akin to the landlord management model can be talked of regarding both economic management and political supervision. 5 The first reform essentially resulted in great managerial autonomy for the various port authorities after Law 27/92, although under the supervision of the central government’s State Port Authority which had to approve their respective business plans on a yearly basis (COTO-MILLÁN, 1996). The second reform entailed political decentralization in Law 62/97. This law established that regional governments should assume competence for naming the presidents of the port authorities. Since then, the vast majority of the presidents have been members of the parties in power in the regional governments where the port authorities are found (CASTILLO-MANZANO et al., 2010). Moreover, coinciding with the port devolution process, a process of over-investment in the Spanish port system similar to that seen in other European countries in the previous decade was also in place (for examples of this, see GOSS, 1995). In this way, the vast majority of the port authorities used their autonomy to make major improvements to their infrastructure or port superstructure. 6 In some cases, such as A Coruña, Ferrol and Gijón, they constructed a new port and moved out of their old facilities, and in others, such as Seville, they constructed a new lock and a new logistics activities zone with its associated quays. In short, investment in the Spanish port system has multiplied by three with this process once inflation has been deducted. To be specific, the €322 million invested in 1991, a year before the port devolution process began, rose to €998 million in 2008 (with both amounts expressed as constant year 2000 euros). It is also easy to see that there is a correlation between the increase in investment expenditure and political decentralization, as up to 1997 investment remained below €400 million constant euros. In this context, many regional governments considered that ports could be turned into major drivers of regional development, with this hypothesis being upheld by a number of studies on port economic impact that were commissioned ad hoc (for a survey of these, see CASTILLO-MANZANO et al., 2004). This led a number of regional governments to take on competences without any legal basis, such as the coordination of the policies and strategies of all the port authorities in their regions. For this, public companies were created in regions such as Andalusia, Catalonia and Galicia that have not been able to undertake this coordination in practice since, as previously stated, it is the central government which has to approve their respective business plans on a yearly basis through the State Ports Authority. In comparison with these new formal organizations, others of a more informal nature have also appeared with the sole objective of lobbying central government in order to attract a greater volume of investment to certain port authorities. To be more precise, these are rather heterogeneous platforms organized within the port authority environments which have gradually incorporated such widely differing organizations as the local councils in the port hinterlands, trade unions, businessmen’s associations and even residents’associations. It has not been uncommon to see representatives of these platforms meeting with representatives of regional or national governments and even from the European Investment Bank. Cooperation between town councils and port authorities has meant additional funds have been secured thanks to the property (real estate) bubble that the Spanish economy has experienced. Specifically, local councils have paved the way for the land where former port facilities were located, which in historical ports was generally in the city centre, to be reclassified, while the port authorities planned their new facilities far from the normally congested urban centres. This reclassification has allowed the port authorities to gain some major extra income by promoting the construction of new facilities (BICHOU and GREY, 2005). The new constructions are used both for company use (offices and conferences) and for leisure activities, in keeping with the Barcelona Port Authority’s successful Port Vell ‘shopping mall and leisure center’model. Port Vell was in turn influenced by successful large-scale international projects such as the major rebuilding project linked to the old docks in London (UK) (regarding this last process, see CARMONA, 2009). Unlike major investment in port infrastructure, there are usually abundant private funds for projects to redevelop old port facilities for tertiary use (malls, offices and even aquariums). Even so, the massive influx of economic resources into ports over the last decade from both the system’s own resources and public resources not linked to the system have not been sufficient to cover these funding needs. 7 This has meant that port authorities have been accruing debt with financial institutions at a doubledigit annual growth rate. In 2008 the accumulated debt stood at over 150% of the total annual revenue for the whole port system. What is striking is the ease with which many port authorities, especially those with little traffic, were able to have access to loans from financial institutions, even 1282 José I. Castillo-Manzano and Xavier Fageda more so when it is considered that after Law 27/1992 they had to respond to present and future loans with their own resources. The only explanation is the belief on the part of financial institutions that any debt would be borne by the central government should the port authority become insolvent. This belief has been confirmed by the Ministry of Public Works’intervention to deal with Gijón Port Authority’s poor debt situation by taking on all its debts in exchange for greater control over its future management. Despite the decentralization described above, the majority of the economic resources needed to fund these investments are still in the hands of the central government in this investment process. This includes public resources not linked to the port system that come from the Ministry of Public Works and from the European Regional Development Fund (ERDF). And the resources of the system itself are also included, specifically both 4% of the income for all ports (except the island ports and Ceuta and Melilla, which only give 2%) as well as resources from an inter-port solidarity fund, known as the Fondo de Compensación Interportuario (Inter-Port Compensation Fund) since 2003. This fund comprises contributions from ports with a surplus to finance investments and other expenditure at ports with a shortfall. These two amounts together come to some €40 million per year, of which €20–25 million are used to finance ports that have special economic difficulties or are located outside the Iberian Peninsula. These amounts are necessary for the economic viability of several of the twenty-eight port authorities in the Spanish port system given their little ability to generate resources, including Ceuta, Marín, Melilla, Motril and Vilagarcía. As such, the issue of how public investment funds are to be distributed is economically relevant both because of the amount in question, €1300 million current year euros in the last decade, which represents an annual average of €46.5 million per year for each of the port authorities, and because of the complicated political interests that are in play. To summarize, it is the central government which authorizes investments and provides a large part of the funding that is required, whereas it is the regional government that can make political capital of them through the president of the port authority, and it is the local councils in the hinterlands of the ports that most benefit from them. For this reason, many mayors of port towns and cities try to front port investment through these platforms, especially in small or medium-sized towns and cities where the port authority is not just one of the leading investors, but the main investor, which means that these towns and cities therefore have great economic dependence on port investment. In this context, this article seeks to analyse the criteria used to split the investments among Spanish ports. The aim is to quantify the influence that transportation economics criteria have on the way investments are shared out and, within this sphere, whether inter-port or inter-regional solidarity criteria prevail or, conversely, it is electioneering criteria to favour cities or regions that are allied with the central government that dominate. The following section estimates an empirical model that accounts for different political and economic factors that may influence the allocation by the central administration of regional investments in ports in the Spanish provinces (NUTS-3) where there are facilities of this type. 8 THE EMPIRICAL MODEL During the period of study covered, the Spanish port system was composed of twenty-seven port authorities which included forty-four ports of general interest. Currently, there are twenty-eight port authorities, after the segregation of the Port of Motril from the Port of Almería. There are other territorial bodies managed by the respective regional administrations, but their size is always quite small compared with the so-called general interest ports. Information on the majority of the variables used in the empirical analysis is only available at the provincial level (NUTS-3), as a result of which the geographical unit of analysis is the province. This means that no distinction can be made between different port authorities in the same province and data for these have been aggregated on the provincial scale. In particular, the sample is based on twenty-two provinces. 9 The time period analysed was also determined by data availability and covers the period 1982–2005. A total of 528 observations were made for the empirical analysis. The empirical analysis is based on the estimation of an equation of the determinants of investments in port infrastructure in province iduring year t. An equation is estimated where the endogenous variable (INVESTMENT) is the amount of investment in port infrastructure in province iduring year texpressed in constant year 2000 euros. Port authorities are in charge of investments in port infrastructure such as berths, while specialized private firms are generally in charge of investments in port superstructure such as cranes. The focus is on the public investment. Data on investment in port infrastructure were taken from historical public capital series that the BBVA-IVIE Foundation publishes periodically. 10 Investments in ports of general interest are proposed by the Spanish Ports Authority (Puertos del Estado), which is an entity dependent on the Ministry of Transport (central government). Proposed investments must be approved in accompanying laws included in the Central Government’s General Budget Law. Thus, regarding the political timing of investments, what is relevant is who is governing at the time at which the How are Investments Allocated in a Publicly Owned Port System? 1283 budget is approved and not when the investment is effectively made. Inthisregard,informationoftheexplanatoryvariables was used referring to province iduring year t–1. Indeed, the levels of investment in year twill be determined by decisions taken in year t–1, whereby the influence of the various explanatory variables is expected to relate to year t–1. The following explanatory variables were considered for the provincial allocation of investments in port infrastructure: (1) The endogenous variable with a lag of 1. Most large-scale port investment projects, from a new terminal to the development of an area for logistics activities, are multi-annual, which means that the investment for period tis probably closely linked with investment during period t–1. (2) The capital stock in port infrastructures in province iin year t–1(STOCK CAPITAL). Capital stock is expressed in constant year 2000 euros and information comes from BBVA-IVIE Foundation. This variable is related to the equity objective of the central government (KEMMERLING and STEPHAN, 2010). If the central government has the objective of equal infrastructure endowment across regions, it should invest more resources in regions with a lower stock of capital. (3) The ratio between the amount of trafficandcapital stock for the port authorities in province iin year t– 2(TRAFFIC/CAPITAL). Traffic data are expressed intermsoftonnesofgoodsmoved. 11 Theinformation on traffic was obtained from the Spanish National Ports Authority, Puertos del Estado. 12 With the data available, this is the best approximation of an indicator of efficiency that the authors have been able to construct for analysing its influence on investments. It should be remembered that previous studies on the regional allocation of investments in transport infrastructure (CASTELLS and SOLÉ-OLLÉ, 2005; SOLÉ-OLLÉ, 2010, 2012; KEMMERLING and STEPHAN, 2002, 2010) used as an indicator of efficiency the ratio of GDP to capital stock. 13 However, the use of infrastructure in relation to capacity is the most accurate indicator of efficiency in the case of transportation. Bearing this in mind, the authors still recognize that their variable has some limitations because it is the ratio between one measure in physical units and another in monetary units. If efficiency considerations are relevant in the regional allocation of investments, the volume of investment received by a port authority can be expected to be dependent upon the volume of traffic in relation to its capacity. The higher the level of trafficinrelationtotheavailablecapacity,the greater the need for investment in improvements and the greater will be the expansion of port facilities. Hence, port authorities that make a more intense use of assets will receive more resources. However,theexistenceof the Inter-Port Compensation Fund could result in the relationship between the levels of traffic and the levels of investment being less closely related than might be anticipated in a system where the financial autonomy of the port authorities is higher than in the case of Spain. 14 It should be noted that port legislation in Spain explicitly states that redistribution based on solidarity depends on the level of traffic that the port generates and is not related to the relative wealth of the province in which the port is located. (4) Gross domestic product per capita of province i during year t–1(GDP PER CAPITA). Information for this variable was obtained from Spain’s Institute of Statistics, the Instituto Nacional de Estadística (INE). Data are expressed in constant year 2000 euros. This variable captures economic wealth so that it measures whether the central government invests more in poorer regions. In such a case, one of the goals of the port policy in Spain would be solidarity based on the levels of income. (5) Port authorities’specialization in containers in province iduring year t–1(WEIGHT CONTAINERS). The information for this variable was obtained from the Spanish National Ports Authority, Puertos del Estado. This variable is constructed as the quotient between tonnes moved in containers by port authorities in province iand total tonnes moved by all port authorities in the Spanish system. Note that several observations do not have container traffic at all. Given that all continuous variables are in logs, the lowest non-zero value in the sample (0.01%) to those observations that do not have container traffic at all is computed. It should be borne in mind that there is generally a more diverse range of commodities in larger ports (DUCRUET et al., 2010a). Controlling for the volume of traffic, this variable is used to examine whether ports that are more highly specialized in container traffic have been favoured with greater amounts of investment. In this way, indications will be obtained as to whether there is a deliberate policy on the part of the central government to favour the development of hub ports. This policy could be one of the reasons that explain why Spain is the European country with most ports on the list of the world’s busiest container seaports, three to be precise. Container traffic is not on average the traffic with the greatest added value (HAEZENDONCK, 2001), but it requires more costly port facilities with a higher degree of technical complexity compared with other types of freight. Furthermore, given the importance of this type of traffic for international trade, it would be plausible to 1284 José I. Castillo-Manzano and Xavier Fageda expect the central government to bias public investments with a view to favouring the creation of one or more hubs on national territory, which would thus make Spain less dependent on other countries for organizing its supply and distribution chains. Finally, and this is probably not just the case in Spain, it should be expected that the greater the value of this variable for a port authority, the greater the presence of private terminals there will be at the port authority (although, generally, with public investment in infrastructure). (6) The percentage represented by added value in manufacturing and construction activities (it was impossible to separate the two activities) compared with the added value that the whole province generates (WEIGHT INDUSTRIAL ACTIVITY). Information for this variable was obtained from Cambridge Econometrics (European Regional Database publication). Manufacturing companies might be anticipated to have a great interest in investments being made in ports where their activities are located as their import and export activities depend on these facilities. Construction companies, meanwhile, could benefit from the construction work that comes from improving and expanding port capacity. Hence, this variable may work as a proxy for lobbying forces. (7) Percentage of votes of the incumbent party in the central government in the elections to the central parliament across regions (VOTES). Information for this variable was obtained from the Ministry of Domestic Affairs website. The party in power in central government might have greater incentives for investing in regions where its popularity is greater as the political capital these regions have for the party will be greater. (8) A dummy variable which takes a value of 1 when the party that is in power in the central government and the party in power in the regional government (NUTS-2) in the province where the respective port authority is located coincide from 1998 (D REGIONAL_CORRESPONDENCE ). Information for this variable was obtained from the Ministry of Domestic Affairs website. This variable is used to measure any political bias there might be in the way that the central government has distributed investments after Law 62/1997 came into force in January 1998. The specific aim is to test the hypothesis that port authorities located in regions governed by the same political party as the central government have been systematically favoured. These include both majority governments and those in coalition. In these regions, the presidents of both the port authorities and of the State Ports Authority would belong to the same political party. (9) A dummy variable which takes a value of 1 when the party that is in power in the central government and the party in power in the largest town or city in the province where the respective port authority is located coincide (D LOCAL_CORRESPONDENCE ). Information for this variable was obtained from the Ministry of Domestic Affairs website. This variable is used toexamine the degree to whichpolitical pressure from platforms that are usually headed by the mayor of the town/city that is most affected by the investment influences central government decisions. (10) Adummyvariablethattakesavalueof1inprovinces where nationalist partiesare in power in the regional government (Basque Country, Catalonia, Canary Islands) during the periods where the party in the central government does not have the absolute majority (D STRENGTH_REGIONAL_PARTIES ). Information for this variable was obtained from the Ministry of Domestic Affairs website. This variable may allow one to capture the pressure of nationalist parties when the party in the central government may need their support in parliamentary votes. (11) A difference in the percentage of votes in the national elections between the incumbent party in the central government and the political party with the second largest number of votes at the national level (SWING) across regions. Previous studies typically use a variable for the difference in a province’s votes between the two main parties in the national elections to test the socalled swing voter hypothesis; regions will receive more investment if the difference in votes between the two main parties is smaller. Information for this variable was obtained from the Ministry of Domestic Affairs website. (12) A ratio which is constructed by dividing the percentage of votes in the province over total national votes in relation to the percentage of seats in the province over total national seats (VOTES/ SEATS). With this variable, the authors wanted to test if the incumbent party in the central government invests more resources in those regions with higher electoral productivity; that is, in those regions where fewer votes are needed to gain seats. Information for this variable was obtained from the Ministry of Domestic Affairs website. ECONOMETRIC ISSUES The data used for estimating the investment determinants equation have a data panel structure. There are various techniques and estimation models available for estimating equations with data panels. Furthermore, the estimates can present heteroskedasticity, non-statio- narity and temporal autocorrelation problems in the error term. Table 1 shows the results of different specification tests. First, it reports tests related to the time series How are Investments Allocated in a Publicly Owned Port System? 1285 properties of the data (stationarity and autocorrelation). The augmented Dickey–Fuller (ADF) test indicates that the dependent variable does not contain a unit root and hence it is confirmed that there is no long-term co-inte- gration relationship. The Wooldridge test for autocorrelation in panel data claims that one may have a problem of serial autocorrelation that must be treated. Second, Table 1 report tests of appropriateness of different estimation techniques in the context of panel data. The Breusch–Pagan Lagrange multiplier (LM) test for random effects shows that random effects are not statistically significant and, hence, a pooled model may be more appropriate than a random-effects model. Furthermore, the Hausman test shows that systematic differences between the random and fixed effects are found and thus the fixed-effects model should be preferred to the random-effects model. Note also that there may be a simultaneous determination of the variables of traffic and investment. 15 In order to deal with this potential bias in the estimation, two different approaches are followed. First, two lags of the variable of the traffic/capital stock ratio are used because the correlation between the first lag and the contemporaneous variable is too high. Secondly, an instrumental variable regression is performed where the following instruments for the two lags of the traffic/capital ratio are used: the third and fourth lag of the same ratio and geographical variables related to location and centrality: longitude and latitude. 16 It would be difficult to maintain that geographical location might be a variable that would explain a bias in the distribution of investment over twenty-four years. On the other hand, geographical location, including the linked attributes of centrality (proximity to origin/destination markets) and intermediacy (insertion in carrier networks), 17 might have an influence on the evolution of a port’s traffic. Thus, for the case of Spain, experience shows that being on a latitude closer to the Straits of Gibraltar brings with it a competitive advantage for bunker traffic, whereas, in more general terms, a longitude in the area of the Mediterranean seems to have a competitive advantage for container transhipment and getting supplies by rail to the great central market of the Iberian peninsula, with the over 6 million inhabitants in the Madrid metropolitan area. Geographical features make this difficult in the area of the Cantabrian Sea. 18 It is argued here that it is important to present the results of different estimation techniques as a robustness check. Hence, the results of the estimation using three different techniques are presented: .A pooled model using the Prais–Winsten regression (panel-corrected standard errors) with an AR-1 disturbance. .An instrumental variables regression where the variable of traffic is considered to be endogenous. .Afixed-effects model with an AR-1 disturbance. All continuous variables are expressed in logarithms. All these estimations include time-fixed effects. Furthermore, they are performed in such a way that standard errors are robust to any heteroskedasticity problem. An advantage of the fixed-effects models is that it allows one to control for any omitted variable which is correlated with the variables of interest and does not change over time. Thus, the fixed-effects model is more reliable than other estimation techniques. A shortcoming of the fixed-effects model is that it may be less informative than other estimation techniques because the within variation of data may be low. Indeed, the political variables of the model only vary each four years as much and the variation rate of the economic variables may be also modest. Note also that different variables in the model may have implications in terms of efficiency, equity and solidarity so that it may be difficult to disentangle all these effects in practice. Hence, different specifications of the equation for the determinants of investments in ports are presented. First, the results considering all variables are presented. Second, the political variables are excluded. Then, the results excluding the variables of stock of capital and GDP per capita are also presented. Finally, the variable of traffic is excluded. RESULTS Tables 2 and 3 present the descriptive statistics and the correlations matrix for the variables used in the empirical analysis. Table 4 reports the results of the estimation of the determinants of investments in Spanish ports when the panel-corrected standard errors with an AR-1 disturbance are used. Table 5 reports the results when an instrumental variables regression is used; Table 1. Specification tests Tests Null hypothesis Results Augmented Dickey–Fuller (ADF) test Dependent variable contains a unit root –0.46*** Wooldridge test for autocorrelation in panel data No first-order autocorrelation 170.90*** Breusch–Pagan Lagrange multiplier (LM) test for random effects No random effect: the pooled regression model is appropriate 2.05 Hausman test No systematic differences in the coefficients of random and fixed-effects models 58.86*** 1286 José I. Castillo-Manzano and Xavier Fageda 9. This specifically affects data for the provinces of ACoruña, Asturias, Cádiz and Pontevedra. The following port authorities are located in A Coruña province: A Coruña and Ferrol-San Cibrao; in Asturias province: Avilés and Gijón; in Cádiz province: Algeciras Bay and Cadíz Bay; and, finally, in Pontevedra province: the port authorities of Marín-Ría de Pontevedra, Vigo and Vilagarcía. Conversely, the Almería-Motril Port Authority straddles two provinces, Almería and Granada. In this case it was decided to assign the observation to Almería province, where the largest of the port facilities are located. 10. For the data, see http://www.ivie.es/. 11. Past traffic is generally considered the most suitable strategy to forecast future traffic through ARIMA models (MEERSMAN, 2005). 12. See http://www.puertos.es/en/index.html/. 13. It is reasonable to use the variable of GDP in relation to the stock of capital as an indicator of efficiency in aggregate studies that place attention on roads and rails. However, it should not be used in the analysis that focuses on specific infrastructures such as ports or airports. For the considered period, roads and rails account for 87.5% of the total stock of capital in transportation infrastructures in Spain. 14. The existence of indivisibilities and economies of scale can also distort the ratio between traffic levels and investment levels, although it is not clear in which direction. On the one hand, a minimum amount of investment is required, whatever the level of traffic, and, on the other hand, highly congested facilities require additional investments that might result in subsequent excess capacity for a certain period of time. 15. In any event, experience in port traffic shows that the endogeneity between traffic and public investments is not so powerful as it might seem a priori, unlike private investments, where if a major shipping line such as the Maerks Line or Hapag-Lloyd spends its own money on a new terminal, its future use by the same shipping line is guaranteed. This guarantee does not exist for public investment, as the use of the new infrastructure will ultimately depend on decisions taken by private and independent companies. For this reason, numerous examples of public over-investment can be found both in- and outside the Spanish port system which have not produced the expected increase in traffic (and this will most probably occur in the medium- or long-term with many of the new ports that are being constructed on the Cantabrian Sea, for example). Furthermore, there may be an endogeneity bias between the political variables and the investment variables. It is difficult to find ways to deal with this problem, which has not been addressed by previous studies on the regional allocation of investments. 16. This is also a way to include the geographical component in the analysis. A weighted average of both variables is included for provinces with more than one port authority based on the volume of annual traffic at each of the port authorities. 17. For a definition of both these concepts, see FLEMING and HAYUTH (1994) and DUCRUET and NOTTEBOOM (2012); and for their application, see DUCRUET et al. (2010b). 18. For a detailed analysis of the importance of the geographical component in modern logistics, see AOYAMA et al. (2006). REFERENCES ALBALATE D., BEL G. and FAGEDA X. (2012) Beyond the efficiency–equity dilemma: centralization as a determinant of government investment in infrastructure, Papers in Regional Science DOI: 10.1111/j.1435-5957.2011.00414.x. AOYAMA Y., RATICK S. and SCHWARZ G. (2006) Organizational dynamics of the U.S. logistics industry: an economic geography perspective, Professional Geographer 58(3), 327–340. BARROS C. (2003) Incentive regulation and efficiency of Portuguese port authorities, Maritime Economics and Logistics 5(1),55–69. BEL G. and FAGEDA X. (2009) Preventing competition because of ‘solidarity’: rhetoric and reality of airport investments in Spain, Applied Economics 41(22), 2853–2865. BICHOU K. and GRAY R. (2005) A critical review of conventional terminology for classifying seaports, Transportation Research Part A 39,75–92. BILOTKACH V. (2010) Political Economy of Infrastructure Investment: Evidence from ARRA Airport Grants. Working Paper (available at: http://www.socsci.uci.edu/~vbilotka/Draft_September10.pdf). BROOKS M. R. and CULLINANE K. (2006) Chapter 1: Introduction, Research in Transportation Economics 17,3–28. CADOT O., HENDRIK-ROLLER L. and STEPHAN A. (2006) Contribution to productivity or pork barrel? The two faces of infrastructure investment, Journal of Public Economics 90, 1133–1153. CARMONA M. (2009) The Isle of Dogs: four development waves, five planning models, twelve plans, thirty-five years, and a renaissance …of sorts, Progress in Planning 71(3),87–151. CASTELLS A. and SOLÉ-OLLÉ A. (2005) The regional allocation of infrastructure investment: the role of equity, efficiency and political factors, European Economic Review 49(5), 1165–1205. CASTILLO-MANZANO J. I. and ASENCIO-FLORES J. P. (2012) Competition between new port governance models on the Iberian Peninsula, Transport Reviews 32(4), 519–537. CASTILLO-MANZANO J. I., CASTRO-NUÑO M., LAXE F. G., LÓPEZ-VALPUESTA L. and ARÉVALO-QUIJADA M. T. (2009) Low-cost port competitiveness index: implementation in the Spanish port system, Marine Policy 33(4), 591–598. CASTILLO-MANZANO J. I., COTO-MILLÁN P., PESQUERA M. A. and LOPEZ-VALPUESTA L. (2004) On the economic impact of ports: a comparative analysis of existing studies for the Spanish portuary system (1992–2000), in Essays of Microeconomics and Industrial Organization, pp. 297–316. Springer, New York, NY. CASTILLO-MANZANO J. I., LÓPEZ-VALPUESTA L. and GONZALEZ LAXE F. (2010) Political coordination costs in the Spanish port devolution process: a note, Ocean and Coastal Management 53(9), 577–580. How are Investments Allocated in a Publicly Owned Port System? 1293 CASTILLO-MANZANO J. I., LÓPEZ-VALPUESTA L. and PÉREZ J. J. (2008) Economic analysis of the Spanish port sector reform during the 1990s, Transportation Research Part A: Policy and Practice 42(8), 1056–1063. COTO-MILLÁN P. (1996) Maritime transport policy in Spain (1974–1995), Transport Policy 3(1–2),37–41. CULLINANE K., SONG D.-W. and GRAY R. (2002) A stochastic frontier model of the efficiency of major container terminals in Asia: assessing the influence of administrative and ownership structures, Transportation Research Part A 36, 743–762. DUCRUET C., KOSTER H. R. A. and VAN DER BEEK D. J. (2010) Commodity variety and seaport performance, Regional Studies 44(9), 1221–1240. DUCRUET C., LEE S. W. and NGK. Y. A. (2010) Centrality and vulnerability in liner shipping networks: revisiting the Northeast Asian port hierarchy, Maritime Policy and Management 37(1),17–36. DUCRUET C. and NOTTEBOOM T. (2012) The worldwide maritime network of container shipping: spatial structure and regional dynamics, Global Networks 12(3), 395–423. ESTACHE A., GONZALEZ M. and TRUJILLO L. (2002) Efficiency gains from port reform and the potential for yardstick competition: lessons from Mexico, World Development 30(4), 545–560. FLEMING D. K. and HAYUTH Y. (1994) Spatial characteristics of transportation hubs: centrality and intermediacy, Journal of Transport Geography 2(1),3–18. GONZÁLEZ M. M. and TRUJILLO L. (2008) Reforms and infrastructure efficiency in Spain’s container ports, Transportation Research Part A 42(1), 243–257. GOSS R. (1995) ¿Una política para los puertos europeos?, Boletín Económico de ICE 2460–2461,5–14. HAEZENDONCK E. (2001) Essays on Strategy Analysis for Seaports. Garant, Louvain. JACOBS W. (2007) Port competition between Los Angeles and Long Beach. An institutional analysis, Tijdschrift voor de economische en sociale geografie98, 360–372. KEMMERLING A. and STEPHAN A. (2002) The contribution of local public infrastructure to private productivity and its political economy: evidence from a panel of large German cities, Public Choice 113(3/4), 403–424. KEMMERLING A. and STEPHAN A. (2010) The determinants of regional transport investment across Europe, in BOSCH N., ESPASA M. and SOLÉ-OLLÉ A. (Eds) The Political Economy of Inter-Regional Fiscal Flows, pp. 297–319. Edward Elgar, Cheltenham. MEERSMAN H. M. A. (2005) Port investments in an uncertain environment, Research in Transportation Economics 13, 279–298. NOTTEBOOM T. (2009) Economic Analysis of the European Seaport System. Report serving as the input for the discussion on the TEN- T policy. European Sea Ports Organization (ESPO), Brussels. SOLÉ-OLLÉ A. (2010) The determinants of the regional allocation of infrastructure investment in Spain, in BOSCH N., ESPASA M. and SOLÉ-OLLÉ A. (Eds) The Political Economy of Inter-Regional Fiscal Flows, pp. 276–296. Edward Elgar, Cheltenham. SOLÉ-OLLÉ A. (2012) Inter-regional redistribution through infrastructure investment: tactical or programmatic?, Public Choice DOI: 10.1007/s11127-011-9896-6 SUYKENS F. and VAN DE VOORDE E. (1998) A quarter a century of port management in Europe: objectives and tools, Maritime Policy and Management 25(3), 251–261. TALLEY W. K. (2009) Port Economics. Routledge, Taylor & Francis Group, New York, NY. TONGZON J. and HENG W. (2005) Port privatization, efficiency and competitiveness: some empirical evidence from container ports (terminals), Transportation Research Part A 39(5), 405–424. VERHOEVEN P. (2009) European ports policy: meeting contemporary governance challenges, Maritime Policy and Management 36(1), 79–101. WANG J. J., NGA. K. and OLIVIER D. (2004) Port governance in China: a review of policies in an era of internationalizing port management practices, Transport Policy 11(3), 237–250. 1294 José I. Castillo-Manzano and Xavier Fageda