Master Thesis
Radboud University
Student: Valeriya Krivda
Student number: 4747771
Supervisor: Dr. Lei Delsen
Second reader: Ludger Buijs
Influence of highway network on economic growth: Empirical evidence from
the former planned economies.
Abstract………..
…………..3
1. Introduction……….…4-6
2. Literature review………....6-13
3. Hypotheses………...13-14
4. Conceptual
framework……….14-15
4.1 Data………....14-15
4.2 Methodology………..15-16
5. Econometric
model………....16-23
6. Results………...23-30
7. Conclusion
……….………...31-32
8. Appendices………33-39
8.1 Appendix A………....33-38
8.2 Appendix B……….
….39
9. References……….40-43
Summary
Countries with the post planned economies have faced various economic performances over the time following the failure of the state-controlled economy system. The target of this research is to shed light on the unstudied differences of the transition economies with respect to highway networking and economic growth. There have been many researches examining the effects of infrastructure on the country development, however this study explores and focuses only on the influence of infrastructural asset. The additional hypothesis implies that a particular distinguishable analysis in the past socialist economies might have an impact of highway network on the GDP. The outcome shows that the highway network has an influence on economic growth but this evidence is quite ambiguous. The results also introduce the possible impact of the country specific effects on highway network-economic development relationship.
1. Introduction
Economists define economic growth as a production function for services and goods, where aggregate economic output is a function of factors of production or production inputs, compared from one time period to another1. Technology, human capital and physical capital
are important for economic growth, especially they play a crucial role in transition economies2.
Transition economy is the one, which has changed its economy from central planning to free markets3. The fall of the Soviet Union in 1991 and its allies in Eastern and Central Europe has
pushed newly created countries to take aim at capitalism and go away from central planning. After around two decades of democratic changes, countries with transition economies are still going through serious difficulties with forming and further improving the sustained economic growth and are still facing problems in elimination poverty and inequality.
These new countries have had an unequally developed infrastructure. The infrastructure is an important component of logistic systems and not developing one has a big impact on economic growth (Vilko and Karandassov, 2011). This has been especially the case for transition economies countries, where transport infrastructure is still poorly developed and is of the main importance (Hilling, 1996; Owen, 1987). That means that the lack of a w ell-developed highway network puts constraints on growth. Even though the relationship between these factors has been admitted, an understanding of their causalities and effects need exploring. (Lakshmanan, 2010). This issue arises due to the lack of information about systems underlying economic change.
Highway networks are a key source of economic growth and development and that fact also has an empirical evidence (Fretz and Gorgas, 2013). Hence, highway networks are always improving for causing the growth. Fernald (1999) provides proof that the building of a domestic highway network in USA between 1950-1960 years triggered efficiency in industries more dependent on vehicles.
Therefore, highway networks affect economy in multiple ways. First, roads are an improvement to transport accessibility, which follows by a development of markets of the individual companies, ceteris paribus, this means an increase in sales with all the positive consequences for the company and its’ employees. It can allow producers to come into 1http://www.investopedia.com/terms/e/economicgrowth.asp, [Accessed 13 Jun. 2017]
2http://philschatz.com/economics-book/contents/m48716.html, [Accessed 13 Jun. 2017]
markets in a cheaper way and to raise the size of their market area (Shatz et al., 2011). Furthermore, it can reduce time within which a supplier can reach markets. Second, the good quality and number of roads means the increasing of mobility of labor resources (Shatz et al., 2011). The employee will be ready to work further from his/her own house with the availability to reach work and home faster. Third, it develops global trade and countries get sustainable competitive advantage that is realized in the flow of investments, more productive conditions that ensure economic growth (Buys et al., 2010).
This paper will focus on physical capital of one of the infrastructure category, the inland transportation (roads, motorways and bridges). Hence, the purpose of this research is to determine to what extent there is a connection between GDP growth and highway assets within a specific country. Meaning a positive influence of the developed road infrastructure on the economic growth rate. This paper will examine the countries of Eastern Europe and those that were effectively a part of the former Soviet Union. I will make an empirical analysis based on the data of six of these countries, which are still in the transition period, due to inaccessibility of needed figures. The sample countries are: Czech Republic, Georgia, Poland, Romania, Russia, and Ukraine from 1995-2015.
There are a lot of empirical studies, which are focused on the encouragement of infrastructure to the growth and level of aggregate output (Sahoo et al., 2010; Sánchez-Robles 1998; Esfahani and Ramirez, 2003, etc). These empirical studies explore the influence of infrastructure (kilometers of electricity generation capacity, kilometers of paved roads, number of telephones, water facilities and etc) on GDP growth in the panel of around 50-75 countries for 40-50 years by using growth model of Barro and Sala-i-Martin (1995) and prove the contribution of infrastructure to economic growth. The empirical research, which study infrastructure and economic growth face with reverse causality issue and deal with it by applying the appropriate instruments. However, since there is a lack of empirical research regarding to the transition economies in Eastern Europe and former Soviet Union, I explore only highway networks. I will present some results of relevant instruments that can correct the endogeneity problem. Moreover, my research is performed using a panel data set of highway network indicators only for former planned economies. Hence, my research question is whether highway networks have an influence on GDP growth in countries with former planned economies.
Section two will present the literature review about contribution of infrastructure and highway networks to economic growth of a country. Section three will introduce the hypothesis I am going to test in further research. Section four will discuss the conceptual framework with data
and methodology parts and section five will present the econometric model, based on the studied literature. Section six will provide and explain the results of the model. Section seven will conclude the results of section six and will give further recommendation.
2.
Literature review
I will distinguish growth theory from empirical evidence in my literature review as well as in my further research. Based on the studied literature, I will follow the framework of the number of papers and extend the models of these previous studies along various dimensions. Infrastructure is one of the source of GDP growth, and it defines the performance of agricultural, service and manufacturing activities.
Several studies on economic growth (Barro and Sala-i-Martin, 1995) consider the role of infrastructure and identify five channels through which infrastructure can positively impact economic growth: as a factor of production, as a complement to other factors, as a stimulus to factor accumulation, as a stimulus to aggregate demand and as a tool of industrial policy (Fedderke and Garlick, 2008). Infrastructure as a direct factor of production contribute to a production process. An increase in accumulated infrastructure capital causes GDP growth. New constructions and maintenance of infrastructure requires investments. Total infrastructure investments are one of the instruments for a country’s economic growth.
The role of infrastructure investments has been studied intensively. The first renowned research in this area was done by David Aschauer. He wrote a number of papers (1989a, 1989b, 1989c) exploring the relations between aggregate productivity output and infrastructure investments. Using the data, he found that public capital stock (like infrastructure of streets, highways, airports, mass transit, sewers, water systems, etc.) is the most important factor determining productivity. Aschauer assumed a correlation between government expenditure on non-military infrastructure involving roads, airports, bridges and economic efficiency. He was studying the impact of public infrastructure assets on United States total factor productivity. He discovered that public capital stock (such as infrastructure of highways, sewers, airports, water systems, streets, mass transit, etc.) is really important element determining productivity by using time-series data. More specifically, he found that a 1% rise in the public capital stock would raise total factor productivity by 0.39%. However, later his papers were criticized for not considering the endogeneity problem.
Moreover, infrastructure investments cause the long-term country economic growth. For instance, Heintz et al., (2009) found, that for the period from 1950 till 1979 growth in
infrastructure contributed almost one-for-one to economic growth. During this time infrastructure investments in areas such as transportation, water management and electricity generation grew on average rate of 4% while the overall economic growth averaged 4.1%. On the other hand, from 1980 to 2007 year growth in infrastructure investments decreased to 2.3%, although average annual GDP growth fell to 2.9% during the same period of time (Heintz et al. 2009). Therefore, the infrastructure investments have lagged behind total economic growth due to a deteriorating infrastructure deficit and necessary investment needs. Besides the contribution from infrastructure investments, there are other valuable sides about countries with well-developed infrastructure. Canning and Bennathan (2004) used an approach to find the benefits of infrastructure by estimating an aggregate production function for a panel of 97 countries over a period of 40 years (1960 to 2000). They included transport infrastructure variable – paved roads, physical capital and human capital as the explanatory variables. The marginal product of infrastructure was measured by growth contribution to the aggregate output. The results show that infrastructure is complementary with human and physical capital, which makes it important for a country’s economic growth. Authors concluded that the rates of return to paved roads are the same or smaller than from the other types of capital in most countries. In small number of countries, they found that there is a lack of paved roads and excess returns to investments in infrastructure. The countries with the lack of paved roads are middle income. These excess returns are a proof of non-optimal investment, which result from a period of sustained economic growth when road construction stocks have lagged behind investments in other forms of capital. This happened due to the small costs of road building, which researchers find in the middle-income countries comparative to richer and poorer countries. These results were obtained from using cointegration methods.
Well-developed countries have different conditions for growth rather than countries with transition economies. Countries with economies in transition have an imperative condition for sustainable economic growth, which is less valuable for already developed countries (Kirkpatrick et al., 2006). This essential requirement is the support of efficient infrastructure assets, such as highway network (Kirkpatrick et al., 2006). Developed infrastructure is necessary to ensure the competitiveness of a country — and indeed to its view of itself and its role in the global community. In turn, competitiveness determines the level of productivity. The level of productivity of a country defines the level of prosperity that can be accomplished by an economy. Furthermore, it can also define the rates of return obtained by investments4. 4http://reports.weforum.org/global-competitiveness-report-2014-2015/methodology/, [Accessed 13 Jun. 2017].
The investments in road infrastructure are still relatively low in countries with transition economies, but very important for their development. Despite this, roads can enhance access to education, jobs and income level of people. Hence, investments in highway network are beneficial for reallocating resources and labour from stagnant sectors of economy to sectors that provide a long-term economic growth. Kaplan and Teufel (2016) show that road network can trigger the formation of underdeveloped countries in five different ways. First, road infrastructure enhances connectivity and the integration between states, ceteris paribus, roads give an access to move quicker and connect with each other. Second, road infrastructure allow enables to travel and communicate cheaper and easier, thus, it improves the effectiveness of provinces across distance. Third, by allocating investments across a country, roads can develop rural areas. Forth, roads increase the ability of government authorities to ensure equal levels of public services across a country. The last channel through which road network can improve a country’s growth is that regional roads develop connectivity between neighboring countries, making it easier to solve different cross-border problems. Countries have to place the priorities that they believe the multiple competing interests and trade-offs and take into account short-term and long-term problems.
Additionally, it is interesting to take a look at the research, which would explore the effect of road network in transition economy within one country. For example, Demurger (2001) used data of 24 Chinese provinces between 1985-1998 and she found that transport highways network infrastructure is the primary source in explaining the growth gaps in the country, by applying growth rates research. Furthermore, she also proved the hypothesis that highway network and investments have a positive and significant impact on economic growth. According to Demurger (2001), determining policy priorities and identifying investments to those that have the highest growth payoff would improve country growth prospects in a long-term perspective. Likewise, expanding and improving highway network would be a useful tool for the development of competitive markets and for the economic growth.
The evidence of how investments in institutional and physical infrastructure contribute to the transition process was brought by Schankerman and Aghion (1999). They conclude that besides the direct cost savings, investments in infrastructure also produce indirect effects. The model indicates that infrastructure strengthen product market competition by reducing transaction costs. This can lead to efficient eradication out of existing high-cost companies. The results indicate that infrastructure investments increase the impetus for comparatively low-cost companies to enter the market and hence, enhance the efficiency of the entry process. These are significantly important processes because infrastructure depends on
elements of the economy, such as the proportion of high-cost firms, the cost of restructuring and the entry costs for new firms, the degree of cost concentration among firms.
Almost the same results were found in the paper of Vilko et al., (2011), where they have figured out that countries from former planned economies withstood a transformation to market orientation and benefited greatly in infrastructural investments. Thus, those countries have a big growth in GDP, since it has been done a lot of investments in transport infrastructure there. In sum, highway network investments, as a variable, can give a positive outcome with regard to transition economies in the final results.
Moreover, in the paper of Estache and Garsous (2012) it is stated that higher quality and quantity of infrastructure can directly raise the growth in transition economies (by ensuring access, roads can (i) improve income levels and jobs, (ii) facilitate private investment, (iii) improve education and markets for farmers’ outputs and others by cutting costs). The above mentioned factors play an important role on how infrastructure raise the economic growth of a country. Due to the limitation of data availability, my research will take into account only quantity of a single infrastructural asset.
Indeed, the higher the number of paved roads the more likely goods will be transported through a country and more people are willing to travel and move along the city/country. Therefore, it shapes its contribution to the economic development of regions, urban and rural areas and it leads to the lower transport costs and final price of goods.
The evidence of this was explored by Badalyan, Herzfeld and Rajcaniova (2014). Their research studied neighbour developing countries (Armenia, Georgia, Turkey), where the prime interest was the impact of transport infrastructure and economic growth. They used the data on rails and roads goods transported (million ton/km), passengers carried (million passenger/km) and the length of both features (km) and transformed all the variables into natural logarithms. The results emphasize the importance of quantity of highway networks and rail/road goods transported in the economic growth process. Besides five ways through which infrastructure can have an impact on the development of a country, which were admitted earlier by Kaplan and Teufel (2016), transport infrastructure influences GDP growth through aggregated demand, which in turn holds an indirect effect. Therefore, this research concludes that there is a bidirectional causality between passengers transported and infrastructure investment, and economic growth and infrastructure investment in short and long run. Since quantity of infrastructure may cause the economic growth in transition economies, I will use the growth rate of total roads and paved roads together with passengers and goods transported to examine its influence on GDP.
The small part of the researches that are concerned with the growth influence of infrastructure, focuses on a single infrastructure asset (Calderon and Serven, 2004). Some studies do this by design, as an example, Fernald (1999) analyzed the productivity outcomes of changes in highway infrastructure in U.S., based on Barro and Sala-i-Martin (1995) growth model, including almost the same variables as previous research (public investments in roads, quantity of roads and total miles driven by truck and auto as a proxy). He also included in his study some explanatory variables as population growth and FDI. His main finding was that the change in roads growth causes disproportional changes in productivity growth in industries with more vehicles. These industries get advantages from road construction, which in turn shows that roads are productive. The paper verifies that constructing a dollar's worth of road adds more than a dollar to economic growth of a country.
There is one more paper, which included FDI in the research. Bhattacharya et al., (2011) was checking sources, which influence the economic growth, merchandise trade and FDI was included in these sources. FDI had a positive significant impact as in the paper of Fernald (1999). Also, they defined merchandise trade as the sum of merchandise export and imports, which and straightforwardly relates to roads and transport. They have found a significance between these two sources and GDP growth.
Besides merchandise trade, one more external factor, which has an impact on economic growth is terms of trade. The Terms of Trade (ToT) takes into account changes in the cost of production and consumption inputs and international demand for countries export (Loayza et al., 2005).
Easterly (2001) claims that such factors as interest rate, developments of technical innovations, debt burden in developing countries explain the situation on markets in transition economies. The change in these factors influence the terms of trade. Loayza et al., (2005) conclude that a country improves its economic growth with an increase of terms of trade, which in turns changed due to the external factors, which are difficult to predict.
Some additional results, regarding highway network and trade within a country, were found by Akpan, 2014. The main aim of the research about South Africa was to determine the impact of length of roads on people’s movements and trading (Akpan, 2014). Akpan clarifies that highway network provides an access to main economic inputs, such as resources, knowledge and technology. For example, the ability of not being stuck to one place and move freely, may bring new ideas about innovations and in turn indirectly stimulate economic growth of a country. Thus, it also reduces the barriers to free movements of people and goods and increases the market for services and goods. Highway network contributes transboundary
trade and investments, increases domestic output and also serves as a stimulus for a country competitiveness. Consequently, using merchandise trade, ToT and FDI will be beneficial for my analysis.
Since some of the countries, which I use in my research are the part of European Union now and some are not, I would like to distinct them as in the study of Falcetti et al. (2005). They proved that countries with transition economies have started to integrate into the world-economy, but those countries, which are the part of the EU nowadays, have made significant advances in economic growth. One of the reasons of such a difference is investments. Some of the studies (Fischer, 1993) have noticed that the degree of the damages from the bad economic policies in the less developed countries is large. Some scientists discovered that differences institutions and governance are decisive for explaining innovations and even the industrial revolution and why modern economic development appeared in the West rather than in other parts of the world (Olson et al., 2000). Therefore, there may be potential differences in results for EU and non-EU countries in my further research.
Among a big variety of papers devoted to infrastructure - GDP relations, the one by Esfahani and Ramirez “Institutions, infrastructure, and economic growth” captured my attention. First of all, this is because the paper “develops a structural model of infrastructure and output growth that specifies the ways in which country characteristics and policies enter the infrastructure-GDP interactions” (Esfahani and Ramirez, 2003). Due to the short time range of a panel data is not possible accurately estimate convergence rates and short-run dynamic indicators with the panel regression method. On this basis, panel data regressions that imply dynamic heterogeneity generate evidence about the allocation of the short-run indicators and help dealing with the biases that can occur if heterogeneity is ignored. They explore the influence of infrastructure on economic output by using data of 75 countries for the period from 1965-1995. They have controlled for terms of trade, population density and investments in infrastructure and used them as exogenous variables. Authors include population density in their analysis because they suppose that it is an important indicator for infrastructure industry since it increases the economies of scale of networks. Having a more concentrated population across a country is likely to make network alignment easier. Nevertheless, this does not have to increase the steady state investment rates due to the population density can decrease the demand for infrastructure per dollar output. Therefore, higher population density can cause stagnation and reduce the rate of alignment to asset imbalances (Esfahani and Ramirez, 2003). Researchers have found that in total the high population density may decrease the growth of a county. Esfahani and Ramirez (2003) show that “cross-country estimates of the
model indicate that the contribution of infrastructure services to GDP is substantial and in general exceeds the cost of provision of those services”. Moreover, authors applied transformed Cobb-Douglas production function with constant return to scale since they have focused on four factors, such as labour, infrastructure and non-infrastructure assets and other factors influence productivity. Each of these factors are broadly interpreted for the reason that any component that may be included in production can be taken as a part of one of these factors. Thus, they take quantity and labour to be exogenous, since they have examined and have proved that Cobb-Douglas function is the only function, which has a direct effect and consistent with a steady-state growth with respect to technological progress that is not solely labour extending (Barro and Sala-i-Martin, 1995). Authors also faced with an endogeneity problem, which may arise in my research as well. They succeeded to find instruments to solve the problem of endogeneity. I will also deal with the same issue and take the possible solutions for it as serial and cross-sectional correlation; measurement error; omitted variables and simultaneous equations.
In contrast to Esfahani and Ramirez (2003) study, I will check highway network infrastructure on the sample of transition economies by differentiating EU and non-EU countries. Therefore, I will follow the same framework of research paper of Esfahani and Ramirez (2003). The further hypothesis will follow from the studied literature above and will be applied into my own research question.
3.
Hypothesis
Transportation and highway infrastructure assets are often mentioned as a key encouraging growth and development.
In his work, Fernald (1999) proved that road infrastructure has an impact on productivity performance. There are some factors, which characterize a highway infrastructural network and through which the highway network has an influence on GDP growth. First, investments in road infrastructure promote to the long run economic growth (Shatz et al., 2011). The support of this argument was proved by Harmatuck (1996), who found the significant and positive effect of investments in road infrastructure on GDP growth. Second, the improvements of highway infrastructure (paved roads and total road’s growth) contribute to the expansion of the trade and intensification of competitiveness of countries and economic regions. This development can occur in parallel with the economy growth, while the growth of the transport sector may generate the growth of employed people and their income.
In order to examine influence of every possible variable highway network includes, I add three hypotheses, which relate to road infrastructure in countries:
H1a: Paved roads’ growth rate has a positive influence on economic growth rate of a country.
H1b: Total roads’ growth rate has a positive influence on economic growth rate of a country.
H1c: The total investments in road infrastructure growth rate has a positive influence on economic growth rate of a country.
Besides the three hypotheses above, such factors as transported goods and carried passengers may also have an impact on economic growth. Based on the earlier studied literature, researchers found out that international and domestic transportations will influence on economic growth. This is a part of a business cycle and the demand side, which has an impact on economic growth of a country. I will correct for these in my further analysis and add two more following hypotheses:
H2: The growth rate of the goods transported on the highways has a positive influence on economic growth rate of a country in the short term.
H3: The growth rate of the transport passengers carried on highways has a positive influence on economic growth rate of a country in the short term.
In addition, I am also interested in the country specific differences that might present it in defining the measure and magnitude of interest by adding a dummy variable for EU and non-EU countries.
My hypothesis is similar to the one that Olson et al (2000) examined in the work namely: ‘‘differences in governance play an indispensable role in explaining why most developing countries fail to grow any faster than the high income countries at the same time that certain other developing countries grow far faster than the rich countries do’’. Despite the studies, where the assumptions on the quality of governance in growing countries was tested, I test my hypothesis on highway network-GDP relationships on a sample of post-planned economy countries. Hence, my fourth hypothesis will test whether there is a difference in economic growth between EU and non-EU countries.
Thus, my last hypothesis is:
H4: There is a positive influence of being an EU member on economic growth than not being in a membership of it.
I will test this hypothesis by using panel data and by making a distinction between countries based on their EU membership.
4. Conceptual framework
4.1 Data
In consequence of data availability limitations, this analysis will take into account the following countries: Czech Republic, Georgia, Poland, Romania, Russia, and Ukraine. From the literature I have been studied, it is also clear that there is a lack of researches about former planned economies with respect to highway network and its influence on GDP.
Nevertheless, it would be interesting to explore more countries with former planned
economies, for example, Belarus, Hungary, Kazakhstan, Latvia, etc. However, some specific variables, such as total roads growth or passengers transported, are absent for above stated countries. Therefore, taking into account more countries would not give me sufficient comparable data. This research shows the effects of highway network on economic development by using data in periods of 1995-2015. Data is obtained from OECD, World Bank database, Eurostat and for some specific variables on National Statistics Offices. I will follow Esfahani and Ramirez (2003) model and the choice of infrastructural measures and include some more variables related to my research based on the studied literature above. The dependent variable in my econometric model is real GDP growth rate. The explanatory variables measure all highway network factors such as: quantity of paved roads, total roads network, investments in roads, goods transported on roads and inland passengers carried (buses, trams, cars). Due to possible high correlation between some variables, I may exclude one of them from my further research. I will also control for FDI, unemployment, population growth, ToT, population density, merchandise trade, EU membership.
All (except EU membership) variables in my research paper will be transformed in the percentage change with the aim to fit to Barro and Sala-i-Martin (1995) econometric model. The highway infrastructure measures are in my main interest.
As I said earlier, for the purpose of checking the effect of joining the EU, I divide countries in two groups and include a dummy variable.
4.2 Methodology
I will estimate the model for highway factors and GDP growth rates of six countries for the selected time framework. Selected theoretical framework (Esfahani and Ramirez, 2003) will
include the use of a panel data regression to test the relationship between GDP growth rate and highway network assets. The model explains GDP growth rate data in a number of countries by a measure of highway variables, and a set of control variables. There is a potential endogeneity problem for the dependent variable or measurement error problems, which may bias the coefficient estimates. Respectively, I will check for the existence of reverse causality by testing the explanatory powers of GDP growth rate and highway
variables. Moreover, I can also face with the effect of business cycle, which may result from the goods transported and passengers carried on the highways. I will undertake the general method of Esfahani and Ramirez (2003) for addressing the endogeneity problem, which is using the lagged values of variables.
5. Econometric model
In this chapter I will represent and find the most suitable approach and adapt it for my
following study. In order to decide on the approach used in this research paper, I searched for the similar studies and adjusted these methods to the peculiarity of my study.
Most studies about the relation between infrastructure and growth followed a model
developed by Barro (1990). I start my research with a Cobb-Douglas production function with constant returns to scale and it takes the following form:
where Y is an aggregate output, α and β are positive parameters, N identifies infrastructural capital, K stands for non-infrastructural capital, L denotes labour and all other factors, Q, that influence productivity (Barro, Sala-i-Martin, 1995). Certainly, N captures all of my focus since it holds the value of my interest. Authors treat Q and L as exogeneous variables. There are some problems arise with regard to infrastructure services. The economies of scale are drawbacks of infrastructure services due to the network externalities, such as poor
performance of government (World Bank, 1994). A second problem is a wide consumption of infrastructure services that makes them politically more receptive (Noll, 1989). These issues are essential for infrastructure segments, because production in those areas is capital intensive and future investors are concerned about the opportunity of ex-post expropriation of their quasi-rents. Hence, the share of infrastructure in output may differ from the indicator α, since
it requires sufficient government intervention. It implies that the small amount of expenditure or cost shares of infrastructure industries in GDP can be deceptive signs of the contribution of those areas to the economy. In general, all the debates about role of infrastructure settles around the claim that β is way bigger than the cost share of infrastructure. To determine β, Esfahani and Ramirez (2003) estimate the production function in log-level form:
Equation 1
where γy ,γk, γn are the growth rates of the per-capita endogenous variables and q is the growth
rate of Q. Nevertheless, this method raises two problems. First one is that non-infrastructure capital stock is tough to measure, because it has to involve all kinds of physical and non-physical capital. Second issue is that infrastructure growth is driven by demand factors that rely on GDP growth. Esfahani and Ramirez (2003) proposed a solution for these issues and they introduced simultaneous equations based on the dynamics produced by variations of the economy from the steady state:
Equation 2 and 3
where units of N and K be an amount of each, that can be produced with one unit of capital, output and infrastructure; δ is the depreciation rate; sn and sk are the shares of output
distributed to the accumulation of infrastructure and capital, respectively. For indication steady-state values by *, authors introduced the following equation:
Equation 4
They imply that the long-term rate of productivity growth, q*, does not change across the countries. The steady-state infrastructure-output ratio, n*/y*, and non-infrastructure capital-output ratio, k*/y, are defined by q*+l+δ also as preferences and the technological and economic factors.
For asset i, i = k, n, using Equations 2 and 4, Esfahani and Ramirez (2003) created a following equation:
where is the gap between the steady state and the initial asset – output ratio. By substituting the equations Esfahani and Ramirez (2003) write a growth equation for the aggregate output per capita:
Equation 6
Since Gk can be shown in terms of variables other than k, researchers transformed it To:
Equation 7
These equations include such variables as Q (productivity factor) and si* (investment-GDP ratio), which cannot be measured directly. Thus, in the further econometric model, they can be substituted by a number of infrastructure variables that act as identifiers. Equation 7 assumes one infrastructure sector. In the paper of Esfahani and Ramirez (2003), they extend this model to a case with more infrastructural industries, ‘‘with each one’s stock entering the production function as a separate Cobb-Douglas factor is straightforward’’ (Esfahani and Ramirez, 2003). Hence, researches end up with the following model, where t and p identifies telephones and power generation capacity, respectively:
Equation 8
Therefore, I presented the short derivation of Barro’s economic growth model as in the paper of Esfahani and Ramirez (2003), which concerns with the same outcome as my econometric model. On the left hand side of the model, I also use GDP growth. On the right hand side besides variables researchers use in their work such as population density, population growth, terms of trade, investments in infrastructure, I add some more control variables such as unemployment growth rate and merchandise trade. Also, I go for the same logarithmic transformation. However, in contrast to Esfahani and Ramirez (2003), I assume one
infrastructure sector in my research and expand it, rather than taking Q and si* and substituting them with multiple infrastructure assets such as telephones and electricity generation
capacity, etc. Hence, I substitute them only with variables, which has a relation to highway network. Their model presents logarithmic transformation with log variables, while in my model all the data is transformed in logarithms already in the beginning. Accordingly to Esfahani and Ramirez (2003), I follow their choice of infrastructural measures and control variables and expand my model more. My econometric model is:
Equation 9
where first five variables define highway network, such as roads paved growth, roads total network growth, investments in roads growth, goods transported and passengers transported growth. ToT is defined as terms of trade and FDI is defined as foreign direct investments. Other control variables are population density growth, unemployment growth, population growth and merchandise trade growth. The last dummy variable is EU membership, which identifies whether a country belongs to EU. The detailed definitions of them are given in Appendix 1.
The dependent variable GDP is one of the key measures of a country’s economic performance. GDP is one of the most common measure of an economy’s production or output. It is determined as the total value of services and goods produced in a country in a specified time period5. Therefore, GDP is a highly accurate indication parameter of an
economy’s size. When any country undergoes economic growth, its average condition of life, the quality of health and education of its population is also supposed to rise.
There are multiple benefits of having a developed transportation infrastructure. It can rise the speed with which producers reach markets and inputs, letting them to keep lower inventories and perform just in time production6. Hence, the possibility of providing wider range of goods
and services with lower costs is increasing. The declining of logistic costs will provide more benefits for consumers and producers. Also, reduced costs can have a benefit because of the increasing of productivity. Moreover, it is simpler to have a benefit of opportunities for investment in human capital, for example, definite advantages (Banerjee et al., 2012). In sum, freedom of people, due to high mobility, may bring new ideas, strivings and information about innovations, which in turn induce the economic growth of a country.
One of my control variables is the percentage change in unemployment growth rate. The unemployment rate is the percentage of the share of labour force that do not have a job. Often it rises or falls due to the changing of economic conditions. For instance, when the economy of a country in a bad shape and there is lack of jobs, the unemployment rate can rise. In case 5http://www.independent.com/news/2014/oct/19/why-gdp-so-important/, [Accessed 12 Jul. 2017].
the economy experiences good time and it grows at a healthy rate and the number of available job places are relatively high, then unemployment rate can fall7. Therefore, I expect to have a
negative sign in front of the unemployment growth rate.
There are multiple benefits of having a developed transportation infrastructure. Besides all the reasons I have mentioned above, regarding to increasing mobility, trade market integration, highway network can also allow producers to have a more extensive choice of input suppliers. It can rise the speed with which producers reach markets and inputs, letting them to keep lower inventories and perform just in time production8. Hence, the possibility of providing
wider range of goods and services with lower costs is increasing. The declining of logistic costs will provide more benefits for consumers and producers. Also, reduced costs can have a benefit because of the increasing of productivity. Moreover, it is simpler to have a benefit from opportunities for investment in human capital, for example, definite advantages
(Banerjee et al., 2012). In sum, freedom of people, due to high mobility, may bring new ideas, strivings and information about innovations, which in turn induce the economic growth of a country.
One more essential factor of GDP is FDI, which in my research is the value of net inflows of direct investments made by nonresident investor in the given economy. The main
characteristic of FDI is that it is an investment made that sets an effective control or at least a significant impact over the decision making of a foreign enterprise9. It also includes assets
and liabilities between direct investment enterprises of a given country and their direct
investors10. A substantial change in GDP growth, often has a big impact on the stock market11.
If the economy goes down, then earning for firms are lower, which converts into lower stock price. Therefore, investors care about negative GDP growth, since it is one of the indicators economists use to define whether a state is in a recession. Moreover, FDI also help countries without a domestic changeover to develop resources, which would not be able to evolve otherwise 12. In sum, FDI has three important advantages, which is an access to markets,
access to resources and reduced costs of production. Therefore, I expect to have a positive
7https://www.frbatlanta.org/-/media/documents/regional-economy/econsouth/11q3fedatissue.pdf, [Accessed 12 Jul. 2017].
8http://www.rand.org/content/dam/rand/pubs/monographs/2011/RAND_MG1049.pdf, [Accessed 28 Jun. 2017].
9http://www.economicswebinstitute.org/glossary/fdi.htm, [Accessed 12 Jul. 2017].
10https://datahelpdesk.worldbank.org/knowledgebase/articles/114954-what-is-the-difference-between-foreign-direct-inve, [Accessed 28 Jun. 2017].
11http://www.economicswebinstitute.org/glossary/fdi.htm, [Accessed 12 Jul. 2017].
12http://www.investopedia.com/articles/investing/030813/look-foreign-direct-investment-trends.asp, [Accessed 28 Jun. 2017].
relationship between FDI and GDP as in the paper of Bhattacharya et al. (2011), where they have found that FDI increase GDP growth.
As Esfahani and Ramirez (2003), I also check terms of trade (ToT). They use terms of trade for identifying productivity shocks and find GDP rises with ToT together. Terms of trade represents the ratio of a certain country’s export prices to import prices of goods. When a country’s ToT enhance, it suggests that for every unit of export that the country sells, it can buy more units of goods that are imported 13. An increase of ToT can indicate the overall
welfare of a country has grown. The change of terms of trade can be influenced by changing in prices due to the rate of inflation and exchange rate. The ratio can be changed according to internal and external factors affecting a certain country, taking into account supply and demand for the goods that are exported and imported. Different factors have an impact on terms of trade. For example, the small amount of goods provided for trade has a dramatic influence on terms. A bigger amount of goods provided for trade means that the bigger amount of suppliers is likely to buy and sell capital gained from sales14. The quality and size
of a good are also factors that affects ToT. The bigger the size of a good, the more profit it is likely to get from selling it. The same applies to the goods with higher quality, which cost more. Moreover, if export relative to import prices grow, it implies that bigger volumes of imports can be purchased with a given amount of exports. This assumes an increase in real purchasing power of domestic manufacturing is tantamount to a reallocation of income from other parts of the world and have higher influence on investments, savings and consumption15.
Consumption is one of the main component of GDP, which affects ToT. If consumption goes up, GDP will also rise. Hence, the terms of trade is important for my further research and I expect to have a positive sign in front of it.
Another valuable factor for my model is population density. Stiglitz and Pleskovic (1999) assume that geography – along with political and economic institutions – still matters for economic growth. A high population density can increase competition between countries. Hence, geographic factor should be taken into consideration in theoretical and econometric research of cross-country economic development. Moreover, Esfahani and Ramirez (2003) argue that population density decreases the cost of network expansion and plays an essential role in enhancing the response to infrastructure gaps. Another similar variable, which goes in line with population density, is population growth. Population growth can worse a country’s 13http://data.worldbank.org/indicator/NY.TTF.GNFS.KN, [Accessed 12 Jul. 2017].
14http://www.investopedia.com/terms/t/terms-of-trade.asp, [Accessed 7 Jul. 2017].
growth. The negative impact of population growth appears to have the most significance in transition economies due to the not well-developed institutions, which cannot provide enough assets and protection to residents. For example, fast population growth in developing
economies has got less advance than probably have been lost opportunities for increasing standards of life. Thus, rapid population growth reduces the time needed to reach the population size that ensure economies of scale in communications, transport infrastructure and production. Moreover, the high population growth can make it difficult for countries to develop administrative structure and human skills that are needed to manage their resources. Hence, there is a high possibility of having a negative impact from population growth and population density on economic development in transition economies.
Following, Olson et al (2000) I also check the role of Merchandise trade, which includes the sum of merchandise export and imports and has a straightforward relation with roads and transport16. Nowadays, the growth of merchandise trade rise rapidly, especially in some
developing economies between 1980-2008 years. This happened mostly due to an increase of world trade openness between countries and financial markets, and globalization of labour migration17. Therefore, export and import can be a stimulus of increasing country’s economic
growth. The more goods are exported, the higher growth of merchandise trade can be.
Nevertheless, globalization can serve as a threat for developing countries. Such threats can be an increased inequality, increased control by developed countries, outflow of labour, etc. Despite all the threats, if a country does not participate in globalization process, it is more likely that it will face with the declining private capital flows and shares of trade and will fall behind their peer group in relative terms (Streeten, 1998). Changes in both exports and imports are good indicators of economic trends in a country. Trade figures have an impact on currency values in foreign exchange market. For instance, currency can be really sensitive to the changes in trade deficit, which a particular country performs, because the trade
imbalances create higher demand for foreign currencies18. The more goods are exported and
imported, the higher the rate of merchandise trade is and its impact on GDP. Hence, I expect a positive relation between merchandise trade and GDP growth, as in the paper of Bhattacharya et al. (2011). A lot of people has an opinion that joining of countries with former planned economies to the European Union has many benefits. Some of such advantages for Eastern Europe of being 16http://data.worldbank.org/indicator/TG.VAL.TOTL.GD.ZS, [Accessed 28 Jun. 2017].
17 https://www.eximbankindia.in/Assets/Dynamic/PDF/Publication-Resources/ResearchPapers/21file.pdf, [Accessed 28 Jun. 2017].
part of EU, for instance, political stability and higher integration amongst European states; non-tariff barriers that is followed by benefits such as more exports for industries with a comparative advantage and lower price for consumers, which in turn creates more
workplaces. An increased internal investment is another benefit, because higher stability in the economy create greater harmonization. Therefore, as already mentioned before, EU membership will be a dummy variable for capturing this effect. I assume that joining EU will have positive effect of GDP growth rate and highway infrastructure.
6. Results
This section presents results that were obtained after running the series of regressions and the issues I have faced with. Around eighty regressions were ran in order to test the hypotheses. I test consistently each of five highway infrastructure variables.
6.1 Results of regressions with lagged dependent variable.
Table 1 below presents the results. The tables are made in such a way that on the top of the regressions results, there is a row, which shows the results of regression it belongs to: 2, 3, 4, 5, 6 stands for panel data models; 2b, 3b, 4b, 5b, 6b corresponds to the models with the lagged value of GDP growth for the same panel data with fixed effects. Model 1 and 1b represent the results of OLS regression analyses. Table 1 includes six models based on different variables, which have been used there.
I ran the OLS regression on the original Model 1 presented in section five. Most of the variables appeared to be insignificant. In the study of Esfahani and Ramirez (2003), they used lagged independent variables as instruments.
In case of these panel data regression models, using lagged independent variables appeared to be insignificant. So, I have addressed lagged GDP growth as my new instrument, since I have assumed that my dependent variable might be determined by its past level. As a result, all lagged GDP growth rate are significant and valid as instruments. Also, the two highway network variables are significant. Paved roads growth and passengers carried growth rate are significant at 5% and 1% confidence level respectively. Both of the variables have a positive influence on GDP growth. Hence, it proves that there is a correlation between them and GDP growth rate. These findings go in line with the previous research and the percentage of variance explained is around 30-35%, which is low. Due to the incomplete and unsatisfactory results, I have got by using OLS regressions, I will apply Panel data.
I have also checked for reverse causality and detected its existence in my panel data regressions. Appendix A5 reports the results on reverse causality. I tested the GDP growth rate on the explanatory variables and one of these tests is reported there. GDP is significant at 1% confidence level, meaning that there is a reverse causality. Nowadays, most of the researchers use lagged dependent variable for correcting the endogeneity problem, I have also applied this method. My dependent variable remained the same, which is GDP growth rate. Therefore, all lagged GDP growth give slightly better results.
Appendix A3 displays the results of the Wu-Hausman test, which generated a Chi-square statistics of 12,85 with a p-value of 0,4598. This means that the null hypothesis of the distinction in coefficients being not systematic is rejected. Hence, the choice between fixed effect and random effect models would not make an important difference. However, I will use fixed effects model, since I expect that being a specific country plays a significant role in the highway network-GDP relations.
Paved roads growth is significant at 10% confidence level in OLS regression with a positive 0.088 coefficient, which means that an increase of 1% of paved roads growth rate will increase the GDP growth rate by 0.088%. This result goes in line with my expectations. Hence, the higher the amount of paved road in countries, the higher the GDP growth in the coming years. However, roads total network growth is not significant in the regression analysis, even though I expect them to be correlated and move in the same direction.
Passengers transported variable has high positive coefficients in all models. It appeared to be significant at 1% confidence level in all regressions. An increase by one percent of passengers transported growth rate, increases GDP growth rate on average by 30%. Passengers transported has a significant influence on economic performance of developing countries. Even though it has a significant influence, the coefficient is really high and detects the probability for biased coefficients due to omitted variable bias. In the next section with the indirect effects regression analysis, I will try to avoid this problem (see results in section 6.2). The other variable, which gives positive and significant results, is the Terms of Trade (Models 1, 2, 2b, 3, 3b, 5, 5b, 6, 6b). It gives the highest impact on GDP growth in the Model 2b with lagged dependent variable. It is significant in all models at 5% confidence level. Taking into account the results, the terms of trade (ToT) has a positive influence on GDP growth rate, which is in compliance with my predictions. The average coefficient of 0.13 testifies that growth of GDP is determined by factors other than the ToT.
Variable model1 model1b model2fe model2bfe model3fe model3bfe RoadPavedGr 0.07603428 .08834237* 0.10018513 0.10457614 RoadsTotNetworkG r -0.0227558 -0.00429565 0.01637116 0.02415623 InvestRoadGr 0.20269289 0.15982429 0.01780783 0.19923343 GoodsTransGr 2.1224693 1.2256521 -2.252348 -1.8997502 -7.5599175 -7.2095069 PassenTransGr 32.040699*** 33.364222*** 31.37809*** 29.862788*** 33.618208*** 28.082*** ToT .06936745** 0.04089404 .17099255** .16930079** .14447743** .12076245** PopDensGr 75.362211 -165.99651 -12.298373 -11.896075 -134.87114 -188.47243 UnemplGrRate 3.1341926 5.8490863 -28.643516 -22.18707 PopGr -3.713546** -4.1146961** -3.7968876 -3.6893932 FDIgr 12.253333 -1.5617193 10.299318 12.391069 54.489912*** 39.877161** EU_member 0.97259101 -0.17794911 -5.1623607** -5.189384** -2.5205071* -1.6990291 MerchTr_Gr -1.9404638 17.798534** 19.747352** 10.957104*** 10.029068***
GDPlevel 1.09E-12 -4.71E-12 -4.35E-12 -4.369e-12* -3.821e-12*
lGDPgr 0.17281228 0.04704647 .21189446**
_cons 3.8531164 -2.2251218 -18.994312** -21.033893** -17.742831*** -15.659363***
N 65 63 65 63 107 102
r2 0.30678627 0.35705652 0.46608493 0.49684184 0.40218259 0.4561421
r2_a 0.16291173 0.16953133 0.25716164 0.27451614 0.31861671 0.36862474
Variable model4fe model4bfe model5fe model5bfe model6fe model6bfe
Table 1. Panel data regressions with lagged dependent variable
RoadPavedGr 0.03821121 0.0342765 RoadsTotNetworkGr 0.02984872 0.02441992 InvestRoadGr GoodsTransGr -7.5599175 -7.2095069 -7.8554433 -7.6530741* PassenTransGr 33.12987*** 28.282325*** 33.618208*** 28.082*** 37.30989*** 32.59552*** ToT 0.07519631 0.07226079 .14447743** .12076245** .1482597** .11897301** PopDensGr -220.92578 243.17204* -134.87114 -188.47243 -159.34688 -183.68679 UnemplGrRate PopGr FDIgr 36.599222* 29.430779 54.489912*** 39.877161** 55.797261*** 37.536712** EU_member -2.9835497 2.2239599 -2.5205071* -1.6990291 -2.7584732* -2.1368645 MerchTr_Gr 10.165313** 9.915549** 10.957104*** 10.029068*** 11.911297*** 10.649238***
GDPlevel -1.52E-12 1.57E-12 -4.369e-12* -3.821e-12* -4.334e-12* -3.666e-12*
lGDPgr .20897134** .21189446** .21383535**
_cons -11.217676** 11.395089** -17.742831*** -15.659363*** -18.851924*** -15.701114***
N 81 79 107 102 104 100
r2 0.39794023 0.46369385 0.40218259 0.4561421 0.43131627 0.4678275
Foreign Direct Investment (FDI) has also a positive and significant impact on the dependent variable, besides first two models. It is mostly significant at 5% confidence level with the lowest coefficient of 36.599and the highest of 55.797. These coefficients are high, which show that the impact of foreign direct investments is important for the country’s economic growth.
Another important result, which come along with my predictions, is the influence of population growth rate. This variable appeared to be significant and have a negative influence on GDP growth, with the coefficients below -4.2. This can happen because population growth is difficult to overcome in the developing countries due to unstable situation with modern institutions, meaning that national institutions are still working not that effective. Such things as poorly developed markets, unproductive government policies and programs fail to ensure a secure protection (Klassen and Lawson, 2007). Moreover, the fast-growing rate of population means a high number of people coming in the labour market, which is not always possible to provide workspaces for everybody. In truth, in developing countries the number of unemployed people is raising so fast that in spite of the effort towards planned development, it is not feasible to provide jobs to all people. Thus, it makes it difficult for economically developing countries to solve their problem and, hence, to support the economic growth. The variable merchandise trade is significant in panel data regressions, which also goes in line with my predictions. An increase of merchandise trade by 1%, increases GDP growth rate from by approximately 10%, which in turn means that trade of goods plays an important role in developing economies and may boost the economic growth of countries.
Another variable, which appeared to have a negative impact on GDP, is EU membership. I did not expect it to have such an impact on economic growth. EU membership is significant at 1% and 5% confidence level. These results are quite surprising since I expected that joining to European Union would improve infrastructure network, increase FDI, exports and, therefore, increase national output. However, the results show that being a member of European Union decreases GDP growth by around 3%. This can happen in a short-term, which is in my case not more than ten years. The economy of a country, which joins the EU may take time to adjust to the new macroeconomics rules, which is not possible in a short run and, hence, have a negative impact on GDP growth rate (also in a short run).
Of course, transportation is vital for the European economy: without good connections Europe will not grow or prosper. A great deal of progress has been made in the last 20 years to improve travel links between the West and the East of Europe. East-West connections that
were completely or partly missing, or restricted to only certain modes of transport, have now been integrated into the new (Trans-European Transport Networks) TEN-T network19.
According to Avery et al. (2009), “one has to be cautious with establishing direct links between EU accession and simultaneous economic, social and political developments. Such links surely do exist, but changes since 2004 cannot be simply seen as results of EU accession. There are other factors – global developments on the one hand, domestic processes on the other – that influence the country’s development and the trends described above”.
However, within the EU, there is still a considerable unbalance in the quality and availability of infrastructure between and within the member states. In particular, according to a European Commission report, the East-West connections require improvement through the creation of new transport infrastructure and/or maintenance, rehabilitation or upgrading of existing infrastructure20.
On the other hand, regional economic integration requires connections. These come from infrastructural links as well as from regulatory convergence and cooperation. Since the accession of new members in 2004 to the EU, and after years of neglect, much effort has been devoted to knitting Central Europe together. Between 2007 and 2013, €36 billion of the €355 billion total of EU cohesion and regional funds was earmarked for the development of roads, railways, ports and airports across the region. Much has been achieved in transport infrastructure although it has been predominantly focused on roads, with rail badly lagging behind. As an example, it is now almost possible to drive from Warsaw to Vienna on a highway but there is no direct highway from Warsaw to Budapest, while the 65km-long train ride between Vienna and Bratislava, the two geographically closest capitals in Europe, still takes one hour21.
Although I expected to have a significance in models for such variables as investments in roads, paved roads, goods transported and population density growth rate, they appeared to have no impact on GDP growth rate in my models.
19 http://europa.eu/rapid/press-release_MEMO-13-897_en.htm, accessed on the 7th of July, 2017
20http://europa.eu/rapid/press-release_MEMO-13-897_en.htm, accessed on the 7th of July, 2017
6.2 Results of regressions with indirect effects and lagged
independent variables.
Table 2 below presents the additional results. Table 2 includes three models based on different variables, interaction terms and lagged independent variables, which have been used there. In order to check for indirect effect, I included interation terms between GDP growth and the roads total network growth; GDP growth and investments in roads; GDP growth and goods carried; GDP growth and passengers transported.
Modelc1 includes results with lagged dependent and independent variables. Modelc2 shows the results with lagged GDP growth rate and interactions terms between the dependent and the independent variables. Modelc3 presents also lagged dependent and independent variables and includes interaction terms between the dependent and the independent variables.
Following the approach of Esfahani and Ramirez (2003), which used lagged values of endogenous variables as instruments, I also added them on the right-hand side. Adding lagged independent variables showed that lagged goods transported and lagged passengers carried gives significant results at 1, 5 and 10% confidence level, serving as good instrumental variables.
Among the interactions variables between the dependent and the independent variables, all four appeared to be significant. Those variables are the interaction terms between the GDP growth rate and the Roads Total network growth rate, the interaction between the GDP growth rate and the Investment in roads growth rate, the interaction between GDP growth rate and Goods transported and the interaction between GDP growth rate and Passengers carried. Consequently, there is some slight indirect effects between GDP growth rate and the four above mentioned variables. The interation variable between the GDP growth and the roads total network has a positive 0.08 coefficient. Thus, correcting for endogeneity and indirect effects let avoid the biased coefficients. For example, the results presented in the Table 1 with regard to the regression coefficient of the passengers transported growth rate, which has a high value of 29, appear to be fake. After correcting for the indirect effect and endogeneity effects, this coefficient decreased from 29 to 10.04 and it is no more significant, confirming our previous statement that the high coefficient of passengers transported growth rate may be due to omitted variable bias.
Table 2. Panel data regressions with indirect effects and lagged independent variables
Variable modelcfe1 modelcfe2 modelcfe3
RoadPavedGr 0.09148787 .09960323* 0.09788055 RoadsTotNetworkGr 0.1446158** -.27186217** -.24096132** InvestRoadGr 0.82535343 -4.4860487* -3.8895689 GoodsTransGr -6.7752588 -4.2285298 -1.2509118 PassenTransGr 10.953652 33.084616*** 10.040624 PopDensGr -36.011026 -22.428045 -64.32023 UnemplGrRate -27.891536 -14.723949 -23.652842 PopGr -1.8088548 -2.4993988 -0.4576435 FDIgr 22.196694 19.013827 20.514222 EU_member -3.2324234 -2.3105151 -1.6893448 MerchTr_Gr 3.883559 5.4476699 -0.17047713 lGDPgr -0.11100864 .15611924* -0.00454162 Lagged independent variables lRTNW -0.2375696 -0.1914963 lInvest 0.23827073 2.1437629 lGoodsTr 22.454673*** 12.442438* lPassTrans 34.558268** 21.761308* lToT -0.01329551 -0.02400179 Interaction variables between the GDPgr and the independent variables
c.GDPgr#c.RoadsTotNer .07734018*** .08237332*** c.GDPgr#c.InvestRoadGr 1.1878943** 1.0703248** c.GDPgr#c.GoodsTransGr 2.2742065* 0.48141059 c.GDPgr#c.PassenTransGr -3.6753358** -0.70174512 _cons 4.0074969 0.1207109 6.7163503 N 58 63 58 r2 0.64723751 0.75710473 0.84067815 r2_a 0.44145938 0.63269496 0.71620796
7. Conclusion
In this paper, I have studied the empirical relations in which highway network influences the process of economics growth along with the impacts of country specific effects. My estimations were based on the panel data on 6 post-communist countries from 1995-2015. According to my results such variables as roads total network and total investments in roads infrastructure have an indirect positive influence in GDP growth rate. Control variables appears to be insignificant in the final (table 2, modelc3) model with using the lagged independent variables and interaction terms between GDP growth and the independent variables.
The higher is the amount of roads are in a country, the less time is needed to move from one place to another. Hence, it increases exports and imports of the country.
In the panel data regressions (table 1), without interaction terms, FDI growth rate, Terms of trade, Population growth rate and Merchandise trade are significant and carry the expected sign.
The research reported in this paper is not free from some shortcomings. The estimated positive influence of highway network on GDP, the confirmation reviewed in this report is decisively mixed. Indeed, I could not prove all five supplementary hypotheses. I managed to confirm H1b. For some causes “the link between infrastructure and growth cannot be precisely quantified for transition countries. One problem with assessing the linkage is the limited statistical data availability and comparability. It is not possible to produce accurate estimations only at a very aggregate level or with very indirect measures. Therefore, one should be very careful when interpreting the results” (Dodonov et al., 2002). Also, after accounting for the reverse causality between highway asset and GDP, the influence of highway network on GDP growth is unsubstantial. It could be because instrumental variables used in the report are different for every model as well as for every highway variable.
EU membership gave me negative and significant influence on a nation’s GDP (table 1). It might be argued that EU accession cannot be simultaneously resulted into economic, social and political developments.
Even though the results proved the existence of country specific effects, parameters of my interest showed some unexplainable outcomes. In reality, countries differ. Therefore, I could not confirm that the more developed highway assets, the higher economic growth of the country.
However, according to my forecasts, highway policy is an essential factor for translating investments into economic growth. Realizing the potential of this effect for economic growth, I would recommend that further research on economic and institutional features that affect steady highway network-GDP ratios should be perused. This means that, as much focus should be paid to the creation of an institutional framework and of regulations as to highway investments themselves.
8. Appendix A1. Variable descriptions (source OECD and World
Bank)
Dependent Variable
- GDP growth (annual %) - Annual percentage growth rate of GDP at market prices
based on constant local currency. Aggregates are based on constant 2005 U.S. dollars. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources.
Independent variables
Highway network variables
- Highway network – all the roads on a country (paved roads and unpaved roads)
- Roads paved growth rate - paved roads are those surfaced with crushed stone (macadam) and hydrocarbon binder or bituminized agents, with concrete, or with cobblestones, as a percentage of all the country's roads, measured in length.
- Total roads network growth rate - total road network includes motorways, highways,
and main or rational roads, secondary or regional roads, and all other roads in a country.
- Infrastructure investments growth in terms of roads - Infrastructure investment
covers spending on new transport construction and the improvement of the existing network. Infrastructure investment is a key determinant of performance in the transport sector. This indicator is measured in euros for the road.
- Goods transported growth - refers to the total movement of goods using inland
transport on a given network. Data are expressed in million tonne-kilometres, which represent the transport of one tone over one kilometer.
- Passengers transported growth – Passenger transport refers to the total movement of
passengers using inland transport on a given network. Data are expressed in million passenger-kilometres, which represent the transport of a passenger for one kilometer.
