Spatial spillovers of economic growth and public spending in Mexico: Evidence from a SpVAR model, 1999–2019

Abstract

The Mexican government has been decreasing its participation as the main investor in the Mexican economy, but it has been increasing its current spending in the different states of the country. This behaviour has influenced the slow economic growth of Mexico in the period 1999–2019. However, the effects of these spending decisions have had a heterogeneous effect on the growth levels of Mexican states. The aim of this paper is to identify these effects by considering the impact that a federal entity generates on its neighbours (push-out effect) and the benefits it receives from its close neighbours (push-in effect) when there are spillovers of this economic growth. The methodology used to analyse the relationship between economic growth and public spending makes use of an econometric model with Spatial Autoregressive Vectors (SpVAR) to test the existing causal relationships. With the SpVAR model we can show the spatial effects of interaction between the thirty-two states that make up the Mexican geography in the period 1999–2019 and to determine spatial spreading and growth causality. The results obtained make it possible to establish that economic growth spreads from one federal entity to another over time and that public spending does not have a fundamental influence on this growth, due to the strong reduction in investment spending.

Introduction

There has been a lot of criticism of the government’s weak role as an investor and job generator in the Mexican economy. This became much more evident with the country’s incorporation into the North American Free Trade Agreement (NAFTA) in 1994 and its modest results in terms of growth; in the 1999–2019 period, the GDP grew at an average annual rate of only 2%. However, this result does not mean that economic dynamism operates in the same way at the sub-national level, since there are states that have grown well above the average; for example, Aguascalientes and Querétaro, which are states in the central region of the country, and they have had very dynamic GDP growth rates of approximately 5% per year on average in recent years.

The heterogeneous growth of Mexican states also reflects a spatially differentiated behaviour of public spending. Although this type of spending is considered to displace private investment in the dominant economic theory, others consider it a complementary and relevant source for attracting new investment (Chang, 1996). The predominant idea in economics that free market forces solve economic problems and that a balanced public spending policy is sufficient to generate the necessary incentives for entrepreneurs to lead growth on their own, which is based on comparative advantages to consolidate industrialization, has not been fulfilled in reality (Devlin and Moguillansky, 2009). In practice, the idea of the State’s non-involvement in the economy has weakened its role as a direct investor in the productive sector and has confined it to being a provider of current spending by providing education, health, infrastructure and security services at the state and municipal level. This has contributed to sacrificing potential for growth in economies such as Mexico’s.

In Mexico, the government has reduced its investment spending, but has increased direct unconditional and conditional transfers at the state and municipal levels. This expenditure is known as branch 28 and 33 (branch 28 and 33) of the public budget, respectively. In Mexico, the federal government makes transfers to states and municipalities through the items of branch 28 and 33 of the public budget, respectively. Those of branch 28 are participations that can be used as desired, while those of branch 33 are contributions allocated to education, health, educational and social infrastructure, and public security. These are expenses that have a greater influence on short-term growth, compromising long-term dynamics at the sub-national level.

This type of spending generates both positive and negative externalities in the states. Therefore, the main objective of this work is to show how the 32 entities that make up the country benefit directly through the allocation of the budget known as branch 28 and 33, but also to show how indirectly they benefit neighbours that surround them through this transfer of resources. This implies proving that the economic growth that an entity obtains through transfers not only stays within the benefited entity but spills over into a set of entities. These spills are higher with the closest neighbours and with those with stronger ties, which can only be detected using spatial techniques such as the SpVAR model. This highlights that subnational interrelationships are essential to strengthen local markets and that supporting the states that generate more spillovers for their neighbours is essential for the aggregate growth of the country.

Therefore, the main objective of this work is to show how the 32 entities that make up the country benefit directly through the allocation of the budget known as branch 28 and 33, but also to show how indirectly they benefit neighbours that surround them through this transfer of resources. This implies proving that the economic growth that an entity obtains through transfers not only stays within the benefited entity but spills over into a number of entities. These spills are higher with the closest neighbours and with those with stronger ties, which can only be detected using spatial techniques such as the SpVAR model. With this we highlight that subnational interrelationships are fundamental to strengthen local markets and that supporting the states that generate more spillovers towards their neighbours is fundamental for the aggregate growth of the country. The latter means that when growth and public spending spill over to a broad set of states, growth is much higher than when it is geographically concentrated. When there are significant spatial spillover effects, national growth can be more accelerated, due to the fact that multiplier spatial effects are generated.

The hypothesis put forward in this text is that the decrease in public spending on investment in states has negative effects on their long-term growth prospects by reducing their productive base, while the increase in current spending has positive impacts only in the short term. This implies that when the largest proportion of government spending is carried out in income distribution items, rather than investment items, a risky and adverse effect is generated on long-term growth and population welfare, which negatively influences their long-term growth and development path.

In summary, this paper starts from the premise that public spending has direct and indirect spatial effects on economic growth, giving rise to positive and/or negative externalities. Not only do public investment decisions directly affect the population living in a particular federal entity (direct impact), but they also influence neighbouring entities (indirect impact). Therefore, in order to target public investment policy adequately, it is essential to know which federal entities generate the greatest economic spillovers in relation to their neighbours. In this context, this paper aims to identify the impact that an entity generates on its neighbours (push-out effect) and the benefits it receives from its close neighbours (push-in effect) when there are economic growth spillovers. The methodology used for this purpose is that of Spatial Autoregressive Vectors (SpVAR) developed by Beenstock and Felsenstein (2007) and Kelejian and Prucha (2004).

In the Mexican context, there are studies on economic growth at the sub-national level that analyse economic growth, convergence and spatial concentration (Esquivel, 1999, Cermeño, 2001, Díaz-Bautista, 2003, Valdivia, 2008, Asuad and Quintana-Romero, 2010, Quintana-Romero et al., 2016, Rodríguez Benavides et al., 2018, Díaz-Dapena et al., 2019, German-Soto et al., 2020, Cabral et al., 2020). These authors, despite showing the trend and/or existence of economic convergence, do not measure or analyse in greater detail the spatial impacts of interaction between Mexican states, nor do they study the role that public spending plays in these processes. With the use of the SpVAR econometric methodology, this paper seeks to overcome the limitations of the aforementioned authors by identifying and quantifying the spillover effect generated by economic growth in Mexican states and their causal relationships.

This paper is divided into five sections. After an introduction contextualizing the topic under study, the second section presents the theoretical discussion on spatial intra-regional relationships and their importance. The third section describes the methodology; nature of the data and spatial methodology. The fourth section presents the estimation results and the main findings of the work. Finally, the last section presents the most relevant conclusions, the policy recommendations arising from the results and the limitations of the research.