Impact of near economic center on the economic growth of neighboring country

Trade expansion allows developing countries to access a large pool of global human capital through contact with foreign firms and markets. This increases domestic research and development (R&D) investment and international R&D spillovers to the economy. Trade also enhances the overall economic productivity through market selection (Coe and Helpman, 1995; Melitz, 2003). Meanwhile, foreign direct investment (FDI) is important for economic development by building infrastructure and basic industries that are essential for economic takeoff. Gross domestic output (GDP) increases as FDI reallocates domestic resources from low-productivity sectors to highly efficient ones. FDI also transfers knowledge and R&D spillovers (Moosa, 2002).

A country can realize the benefits of trade and FDI more easily if it has a prospering neighboring country. This is because the trade volume between countries is proportional to their economic size and negatively related to the distance between them, as suggested by the gravity theory of international trade.

The benefits of being close to a prospering country are not limited to direct economic exchanges such as trade and FDI. A country can adopt successful development strategies and institutions more easily if it is closer to a successful country than if it is farther away. These include export-promoting strategies and various institutional factors known to be important for economic growth, including educational, administrative, legal and political systems.

This study investigates the hypothesis that the affluence of a country flows into neighboring countries, changing the trajectory of economic growth of those countries. Specifically, this study attempts to estimate a neighboring country’s impact on economic growth in Asia by constructing a unique dataset of the near economic center (NEC) of a country and combining it with the Penn World Tables database. It identifies the NEC of each country based on trade and FDI data that change over the years as economic geography changes continuously. We select the most influential country for each observation as the NEC. Researchers have not explicitly estimated the neighboring benefits of NEC in explaining Asian growth. To the best of our knowledge, no other study has incorporated the impact of NEC on the growth of neighboring countries in Asia in the context of spatial growth regression.

In the economic growth literature, a group of countries that grow together following a leading country is known as the flying geese model. This model explains the economic growth of East Asian countries, including Chinese Hong Kong, Singapore, Taiwan and South Korea, which have followed Japan’s development model to industrialize their economies. This model also explains the economic growth of China and Southeast Asian countries that took the same path as the four countries in accelerating their growth. The model suggests that a developed country transmits economic prosperity to neighboring developing countries by demonstrating a successful growth model, promoting industry development through inter-industry trade and enhancing specialization through intra-industry trade and FDI (for details, see Kojima, 2000; Ozawa, 2010).

In this regard, this study provides empirical evidence for the flying geese model by estimating the growth transferred from a prospering country represented by NEC to neighboring countries. The existence of growth transmission would confirm the pattern of economic growth illustrated by the model in which a group of countries follows a leading country. Although the pattern of Asian growth as described in the model is well accepted and visualized, it is hard to find empirical support based on spatial growth regression. This lack of research is surprising, considering a large body of literature that shows the existence of spatial externality in economic growth (for a survey, see Abreu et al., 2005; Anselin, 2010). This study estimates the impact of growth transmitted from NECs to neighboring countries in boosting the growth of developing countries in Asia.

Furthermore, this study attempts to identify the factors of developing countries that facilitate the successful absorption of neighboring effects delivered through spatial externality. This study considers how factors such as openness, cultural aspects, political stability and growth strategy intervene in activating neighboring effects. This is to investigate the mechanism by which a country is related to NEC by using a set of variables that interact with the economic growth of NEC.

Previous spatial studies investigated the factors of a country that cause spatial externality, including technology, human capital, R&D capital, renewable energy, good institutions, globalization, public investment and infrastructure (Ertur and Koch, 2007; Azorin and Vega, 2015; Tsangaris et al., 2024; Schembri et al., 2024; Ahmad and Hall, 2017; Jayadevan et al., 2024; Alamá-Sabater and Cantavella, 2019; Cohen, 2010). These studies revealed various factors that affect the growth of neighboring countries through spatial externality. However, there are no studies that explore the factors that facilitate the absorption of neighboring effects delivered from NEC although the economic condition of a country determines the actual realization of neighboring benefits. From this, this study can explain the reasons why the economic prosperity of NEC gravitates to only a few countries that fly following a leading prospering country nearby.

Estimations of this study show that there is a significant premium in economic growth for having a prospering neighboring country. Specifically, a country’s output growth increases by about 0.14% if the NEC output growth increases by one percent. This study suggests that Asian growth significantly benefited from a developed country that transmits economic prosperity to neighboring developing countries.

Growth studies investigated the spillover of growth between neighboring countries, acknowledging that economies are spatially dependent. These studies showed that the economic growth of a country transfers to neighboring countries through spatial externality. Most of the studies estimated a positively significant impact of the economic growth of neighboring countries on the growth of a country. However, some studies reported that neighboring effects are insignificant or negligible. Estimates of neighboring effects vary depending on samples and model specifications.  Table A1 ^[1] presents a summary of previous studies that estimated neighboring effects.

Many of the previous studies have investigated the spillover of growth between countries by using European samples. This is because European countries, especially EU countries, are integrated, which is conducive to spatial externality (For example, Amidi and Majidi, 2020; Arbia et al., 2010; Benos et al., 2015; Carrington, 2003; Dapena et al., 2019; Ertur et al., 2006; Fiaschi and Lavezzi, 2007; Fingleton and Lòpez-Bazo, 2007; Fischer et al., 2009; Le Gallo and Dall’Erba, 2008; Le Gallo and Ertur, 2003; Le Gallo et al., 2005; Piras and Arbia, 2007; Ramajo et al., 2008). Some of the studies investigated spatial externality by using world samples (For example, Ahmad and Hall, 2017; Ertur and Koch, 2007; Ho et al., 2013; Jayadevan et al., 2024; Kim, 2024; Keller, 2002; Moreno and Trehan, 1997; Ramirez and Loboguerrero, 2002; Sanso-Navarro et al., 2020; Weinhold, 2002).

Most of these studies showed that neighboring effects are the significant factors that affect the growth of a country although their estimate varies across studies, ranging from 0.11–1.09%. This suggests that the growth of a country increases by 0.11–1.09% if its neighboring country’s growth increases by 1%. However, Fiaschi and Lavezzi (2007) showed that neighboring effects become insignificant when a set of control variables that affect the growth of a country are included along with a country dummy. Jayadevan et al. (2024) reported neighboring effects are negatively significant in the spatial correlation approach that considers a narrowly defined neighborhood.

Noticeably, estimates of neighboring effects depend on the formation of spatial weight that restricts spatial externality originating from neighboring countries. Spatial weight is usually represented by the geographical distance from one country to another in a decaying order. This is because distance functions as an obstacle to the transmission of economic growth between countries. In this formation, neighboring effects seem to decrease if neighboring countries include a larger number of countries. For example, Le Gallo and Ertur (2003) estimated neighboring effects as 0.42% by restricting 10 nearest regions as neighbors in a European sample of regions. Sanso-Navarro et al. (2020) reported much greater neighboring effects of about 0.75% by restricting 5 nearest regions as neighbors in a world sample of regions.

Distance cannot represent non-geographical factors that determine economic proximity between countries, such as bilateral trade flows and institutions. Neighboring effects change if geographical distance is replaced by other measures of proximity between countries. Obviously, bilateral trade flows can be used as a proxy to denote economic distance between countries (Amidi and Majidi, 2020; Weinhold, 2002; Ho et al., 2013). The proxy is expected to capture spatial externality better than distance because technology and economic growth are transferred through international trade. Amidi and Majidi (2020) and Ho et al. (2013) showed that trade affects growth of neighboring countries through spatial externality by using a sample of European and OECD countries, respectively. According to Ho et al. (2013), the growth of a country is affected not by geographically nearby countries but by trade partners.

Trade and distance are closely related as geographic distance determines trade flows between countries. It is necessary to separate the effect of geographic distance on trade from total trade volume to accurately estimate spatial externality delivered from trade relationships. In this regard, Weinhold (2002) estimated trade flows unexplained by geographical distance, size and other cultural links by using a gravity model. The study reported that neighboring effects are significant (insignificant) for developing (developed) countries by using a sample of world countries. This suggests that technological spillovers delivered through bilateral trade are significant for developing countries that should catch up with the technological frontiers of developed countries.

Some researchers denoted economic proximity by economic factors other than trade flows. Benos et al. (2015) defined economic and technological proximity in terms of GDP per capita and R&D output, respectively. This study showed that interregional externalities matter for European regions, regardless of the way neighborliness is defined. Weinhold (2002) used manufacturing trade unexplained by geographical distance.

Another interesting development is using institutional proximity between countries to capture spatial externality in economic growth. The closer the countries are connected by sharing the same institutions, the greater the spatial externality in growth gets. Arbia et al. (2010) showed that neighboring effects increase from 0.58–0.71% to 0.74–0.78% when geographical proximity is replaced by geo-institutional proximity. Ahmad and Hall (2017) estimated neighboring effects as 0.11–0.13% and 0.16–0.21% after denoting institutional proximity by property rights and political institutions, respectively. These results suggest that good institutions promote the growth of a country, which in turn generates positive spillover effects on neighbors’ growth.

Previous studies used different restrictions in defining neighboring countries that can affect the growth of a country. Countries located far away in terms of geographical or economic proximity do not have a significant influence on the growth of a country. There must be a certain criterion about the countries included as neighboring countries. Spatial weight becomes zero for excluded countries, and there is no growth externality from these countries. Many studies defined neighboring countries by simplifying restrictions such as 5 or 10 nearest countries (e.g. Dapena et al., 2019; Le Gallo and Ertur, 2003; Sanso-Navarro et al., 2020; Weinhold, 2002), and countries within a first or second quartile distance or a certain fixed distance (e.g. Ertur et al., 2006; Fiaschi and Lavezzi, 2007; Le Gallo and Dall’Erba, 2008; Keller, 2002).

In contrast, other studies include all the sample countries as neighbors to each other without any cut-off (e.g. Ertur and Koch, 2007; Ho et al., 2013; Ramajo et al., 2008). This inclusive approach helps researchers avoid setting an arbitrary restriction about neighboring countries although spatial externality becomes vague. Other studies divided sample countries into groups or clubs depending on various characteristics such as sharing a free trade agreement, the same area and contiguous border. Countries in the same club become neighbors to each other (Carrington, 2003; Ciccone, 1996; Fischer et al., 2009; Kim, 2024; Rietveld and Wintershoven, 1998; Weinhold, 2002). This narrow approach considers spatial externality, resulting from direct interaction between countries connected closely either geographically or economically. This is useful to investigate neighboring effects that a country can enjoy by joining a free trade agreement or economic union (Kim, 2024).

Transfer of growth between neighboring countries affects regional economic convergence, and growth studies incorporated spatial externality in the regional convergence model. Studies have found an increase in the speed of convergence when it is augmented with neighboring effects. For example, Ahmad and Hall (2017) reported that the speed of convergence increases from 0.8–0.9% to 1.9–2.2% after including neighboring effects. Ho et al. (2013) suggested that the speed of convergence increases by 4.1% with spatial effects through trade linkage.

Studies showed that convergence follows a separate trend across regions and sectors. Carrington (2003) suggested that the speed of convergence is greater within a region than across regions, and regions diverge almost at the same speed in the EU. Dapena et al. (2019) also found different convergence trends depending on the spatial level of observation in which international convergence coexists with intranational divergency in the EU. Ramajo et al. (2008) found that regions in the EU cohesion-fund countries (Ireland, Greece, Portugal and Spain) converge much faster than the rest of the regions of the EU. Le Gallo and Dall’Erba (2008) reported the existence of heterogenous convergence at the disaggregated sectoral level in the European regions.

Previous studies utilized various samples to estimate neighboring effects and their implication on regional growth. Despite a large number of studies, there are no studies that explored spatial externality by using an Asian sample ^[2]. To fill this research vacuum, this study attempts to estimate a neighboring country’s impact on economic growth in Asia by constructing a unique dataset of the NEC of a country.

Our study investigates growth spillover resulting from a single neighboring country tied economically most closely to a country in Asia. Our study follows the previous studies that confined spatial externality narrowly (Carrington, 2003; Ciccone, 1996; Fischer et al., 2009; Rietveld and Wintershoven, 1998). These studies confined neighboring countries to contiguous countries. A neighboring country in our study is the single most influential country as the largest trade partner and FDI country of a country. In this sense, we label our neighboring country as the NEC. From this, we can estimate neighboring effects transferred from NECs to neighboring countries. This fits well with modelling Asian economic growth radiated from NECs to neighbors as observed from East and Southeast Asian growth.

Previous spatial studies investigated the factors of a country that cause spatial externality including technology (Calvo et al., 1996; Ciccone, 1996; Ertur and Koch, 2007; Keller, 2002), human capital (Azorin and Vega, 2015; Cuaresma and Mishra, 2011), R&D capital (Tsangaris et al., 2024), renewable energy (Schembri et al., 2024), good institutions (Ahmad and Hall, 2017; Arbia et al., 2010), globalization (Jayadevan et al., 2024), public investment (Alamá-Sabater and Cantavella, 2019) and infrastructure (Cohen, 2010). These studies revealed various factors that affect the growth of neighboring countries through spatial externality.

However, there are no studies that explore the factors that facilitate the absorption of neighboring effects delivered from neighbors. This is surprising considering that many previous studies have shown that convergence follows a separate trend across regions and sectors. It is observed that the economic prosperity of NEC gravitates to only a small number of countries with a good economic environment that facilitates the actual realization of neighboring benefits. This study investigates the reasons why only a small number of countries can fly following a leading prospering country nearby.

This study compiles the NEC variable by finding a nearby country that has a strong economic relationship. The largest FDI investment country is selected as the NEC of a country although both FDI and bilateral trade flows are considered. Owing to data availability, especially for FDI inflows, NEC is represented by trade volumes for many countries. In determining NEC, we also checked whether countries are strongly connected by FTA and other political relationships. Economic relationships between countries have changed over the years. Thus, this study identifies the NEC of a country for each period by identifying the changing dynamics of economic relationships for each country over the sample years.

After identifying the NEC for each country and year, they were merged into the Penn World Tables 10 to produce a dataset for empirical application. Based on the database, a panel of 49 Asian countries for 1950–2019 is compiled.  Table A2 ^[1] presents the definitions and construction of the variables used in the estimation.

To visually check the relationship between the economic growth of a country and its NEC and to ensure that it is not driven by outliers,  Figure A1 ^[1] presents an added-variable plot to a panel fixed-effects model. The plots are based on the estimations of the fixed effect panel model run against a set of variables that are generally used in growth accounting, along with openness and NEC growth; the included variables are the same as those in Model 1 in  Table A4 ^[1]. The plots show the correlation between a country’s growth and its NEC growth, conditional on other independent variables.

The plots show a clear positive correlation between economic growth and NEC, and the partial correlation is statistically significant. There are some observations in which the growth rates are stable, but the growth rate of their NEC changes significantly. These samples include countries such as Jordan and Syria, which have Iraq as the NEC, and the two Iraq wars that cause these uneven noises. After eliminating the outliers, the correlation became stronger, as observed from the change in  Figure A1.A ^[1] to  Figure A1.B ^[1]. Although eliminating outliers improves estimation results, this study continues to use the total sample. This is because the study attempts to estimate the economic impact of a neighbor, regardless of whether it is good or bad, and under a boom or bust. This also eliminates any arbitrariness in sample selection, which may cause sampling bias.

The economic impact of NEC’s growth on its neighbors looks positive and significant. Now, we investigate this relationship further by adopting an adequate regression model that can deal with endogeneity intrinsic to the growth regression. In addition, this relationship should be replicated by including controls that affect economic growth.

We implemented Pesaran’s (2003) panel unit root test in heterogenous panels with cross-section dependence. Test results show that all the variables are stationary at first difference with I(1) although test statistics are slightly different depending on regression specification (see  Table A3 ^[1]). Thus, we will use growth rates denoted as a log-difference of the variables in the estimation.

To choose an appropriate model for our dataset, we investigate various panel models using a basic model, which includes employment, capital, total factor productivity (TFP) and openness to explain a country’s GDP growth. The basic model also includes the GDP growth of NEC, which is the variable of our focus.

The regression model is expressed using the following growth equation:

where $∆yit$⁠, $∆llt,$$∆kkt$⁠, $∆TFPit$ and $∆yitNEC$ are the growth rate of real output, employment, capital stock, TFP and NEC’s real output for country i at period t, respectively. Open represents openness of a country as denoted by an average share of export and imports in its GDP, and $εit$ is the error term. In this setup, the coefficient of NEC growth, $∆yitNEC$⁠, captures the growth owing to NEC unexplained by traditional growth accounting. Thus, it denotes the premium for having a growing economy as an NEC.

We estimated various models to find the appropriate model for our dataset. Based on model diagnostics, we chose the GMM-FE model because it fits the data well and addresses endogeneity, which is our biggest concern in this growth regression ^[3]. The model is also flexible because it includes dummy variables as additional regressors, such as the time dummy, to eliminate common business cycle effects that simultaneously affect the country (see  Table A4 ^[1]).

Notably, NEC’s GDP growth has a significantly positive impact on a country’s GDP growth, regardless of the model. Estimations from the GMM-FE model show that the GDP of a country increases by approximately 0.141% as its NEC increases by one percent.

Now, we investigate whether NECs’ economic influence on neighboring countries holds by adding various control variables that account for a country’s GDP growth.

The regression model is as follows:

where Z are additional control variables. Note that the equation is estimated using the panel instrument variable two-step GMM fixed effect model, and $β1i$ is the country fixed effect.

To eliminate the cyclical shock that affected the Asian economies simultaneously, this study includes two economic crises in the growth equation: the Asian financial crisis of 1997–98 and the world mortgage crisis of 2007–08 in the growth equation. The estimations show that the Asian financial crisis had significant negative effects on the Asian countries’ economic growth (see Model 5 in Table 1). However, the global mortgage crisis did not have a significant effect.

To investigate the impact of trade and openness on economic growth, the estimation includes openness, measured as the average of import and export shares in GDP. The proxy is stationary and not a growth variable, like the other explanatory variables used in the model diagnostics [⁴]. However, the variable is transformed into a logarithm, and the estimation results are reported in Model 6 in Table 1. The coefficient becomes insignificant after the log transformation. Thus, the study retains the original term, which is denoted as a ratio like any other variable in the model diagnostics. To dig deeper, this study separately inputs exports and imports into the estimation. The coefficient estimates are significant and positive for both trade variables.

This study includes political characteristics such as the communist and transition economies in an estimation to account for the wide political spectrum of Asian countries. As expected, the communist economy has a significantly negative impact on economic growth, but the transitioning economy does not have a significant impact. It is likely that some transition economies will deal better with NECs during the transition period than others do.

Throughout the exercise, the impact of NEC’s GDP growth remained significantly positive and stable, with estimates ranging from 0.136 to 0.142. This confirms the existence of a salutary economic impact running from the NEC to its neighboring economy. Building on these findings, this study investigates how NEC’s growth impacts neighboring economies.

To investigate the mechanism by which a country is affected by NEC, this study uses a set of variables that could interact with the economic growth of NEC. This study individually inputs these variables to avoid possible multicollinearity.

The regression model is as follows:

where $kit∆yitNEC$ is an interaction term included to capture the impact of factor k that facilitates the delivery of NEC’s growth to its neighboring country. If the sign of its coefficient $δ$ is positive, the factor enhances the impact of NEC’s growth on the neighboring country.

First, it is natural for the impact to be delivered quickly and easily if a country is an open economy that is integrated with the world economy. Contrary to this presumption, the estimations show that NEC’s effects do not interact with either the exports or imports of a country. This might reflect the fact that an open economy has economic exchanges with many other economies far and wide in the world, rather than its interaction with an NEC.

However, the impact is expected to increase if direct exchange between the two countries increases. To investigate this, the existence of an FTA between the two countries is used to capture the strength of the economic exchange between a country and its NEC. Surprisingly, the FTA significantly reduces the NEC’s economic influence on its neighboring countries (see Model 12 in  Table A5 ^[1]).

This finding confirms the interpretation that an open growing economy does not solely depend on its NEC growth. A growing economy benefits many countries other than its NEC. For example, the four East Asian Tigers had Japan as their NEC until China overtook it; their economic growth has been organized toward export-promoted growth, but their most important trade partner is the US and not Japan, even though these economies emulated the Japanese economic success and were influenced by its FDI and trade. This characteristic also accounts for the economic growth of the newly emerging Southeast Asian economies since the 1990s.

The distance between a country and its NEC affects the latter’s influence on the economic growth of the former. Trade costs increase, hindering the economic exchange with which NEC exerts its impact as countries move farther away from their NEC. However, the distance between a country and its NEC reduces the delivery of NEC’s growth to neighboring countries, but the effect is insignificant.

Economic exchange increases as countries come closer. However, our estimations show that the distance does not significantly affect the economic impact of an NEC on its neighboring countries. This reflects a well-developed trade route in Asia that connects various countries to their NECs such as Japan and China directly, which are the two most frequent NECs. This coincides with the geographical location of the fast growing or developed Asian countries along the coastlines of the Southeast and Northeast Asia. In addition, this study considers the influence of an NEC and not an economic power that is far away. Thus, distance does not matter for countries within the same region.

Another geographical feature of a country, landlockedness, is expected to interact negatively with NECs. Contrary to this presumption, it enhances the delivery of NEC’s economic growth to a neighboring country. This is because a landlocked country does not have as active an economic relationship as other countries with open trade ports, and thus, it naturally depends much more on its NECs.

Thus, this result is in line with previous estimation results that an FTA negatively interacts with NEC’s economic influence, and that exports and imports do not enhance NEC’s influence. A landlocked country relies on its NEC, whereas an open economy, whether geographically or economically, does not.

Therefore, two economic regimes, the communist and transitioning economies, interact with NEC’s economic growth to see if these factors intervene in transmitting NEC’s growth to neighboring countries. Estimations show that a transitioning economy enhances the impact of NEC’s influence on its neighboring countries, but a communist economy does not significantly enhance it. A transitioning economy requires substantial economic help from the outside, and its economic status helps bring in or welcome NEC’s influence on the economy. However, a communist economy does not significantly enhance NECs influence on its neighboring countries, even if it is positive.

Finally, NEC growth interacts with a dummy variable denoting that a country uses the same language as its NEC. Estimations show that language is neither a barrier nor an enhancer of NEC’s growth. This results from the fact that there are as many countries growing fast together with the NEC without sharing the same language as the other countries.

Estimations of this study show that there is a significant premium in economic growth for having a prospering neighboring country. Specifically, a country’s GDP growth increases by 0.136–0.142% when NEC’s GDP grow by 1% in a simple growth regression. The growth of a country increases by about 0.14% after adding various factors that intervene in the delivery of NEC’s growth to a neighboring country.

Regarding the intervening factors, a country’s distance from NEC, openness (either exports or imports) and use of the same language do not significantly facilitate the economic delivery of NEC to its neighboring country. The transitioning economy and landlockedness significantly increase the impact of NEC’s growth on its neighboring countries, but the communist economy significantly reduces the spillover impact of growth.

Interestingly, the existence of an FTA between a country and its NEC significantly reduces the impact of NEC’s growth on neighboring countries. A country with an FTA with its NEC is likely to have free trade agreements with many other countries. In other words, the existence of the FTA represents the openness of the country. Thus, estimations suggest that a country depends less on a single NEC as it becomes more open. This is in line with the estimations of openness that an increase in exports and imports does not enhance NECs economic impact on neighboring countries.

This study empirically supports the flying geese model, in which a group of countries grow together following a leading country. The model suggests that a developed country transmits economic prosperity to its neighboring developing countries. In this regard, this study supports the model by providing empirical evidence that the growth of a prospering country is transferred to a neighboring country.

This study suggests that a country should maintain a good relationship with neighboring countries to absorb their growth through spatial externality. A way to economic prosperity should start with enhancing economic relationships with neighbors. For this, countries should avoid an unnecessary dispute that often escalates to a trade conflict, which blocks the free flow of spatial externality between countries.

Future studies can check if our results hold when output growth is replaced with total factor productivity growth. This is to investigate spillovers of productivity growth between neighboring countries. Productivity growth is the significant factor that sustains the economic growth of a country. Thus, a new study can explore whether spatial externality affects the sustainability of a country’s economic growth.

The author thanks the editor, Hoai Nguyen, an associate editor and two referees who provided comments anonymously. Their comments greatly helped to improve an earlier draft of the paper. This work was supported by The Japan Society for the Promotion of Science [JSPS KAKENHI Grant Number 22K01492].

The supplementary material for this article can be found online.