• Editorial Board +
• For Contributors +
• Journal Search +
Journal Search Engine
ISSN : 1598-7248 (Print)
ISSN : 2234-6473 (Online)
Industrial Engineering & Management Systems Vol.20 No.4 pp.720-731
DOI : https://doi.org/10.7232/iems.2021.20.4.720

# Determinants of Export Diversification in Developing Countries

Rossanto Dwi Handoyo*, Solihin, Kabiru Hannafi Ibrahim
Faculty of Economics and Business, Airlangga University, Surabaya, Indonesia
Faculty of Social and Management Sciences, Federal University, Birnin Kebbi, Nigeria
*Corresponding Author, E-mail: rossanto_dh@feb.unair.ac.id
July 3, 2021 October 25, 2021 November 25, 2021

## ABSTRACT

This study unravels the impact of different determinants of export diversification in 62 developing countries classified as low, middle, and high-income over the period 2010-2018. The empirical strategy based on Poisson Pseudo- Maximum Likelihood revealed that GDP promotes the diversification of export in low, middle, and all countries’ samples and reduces it in high-income countries. Human capital reduces export diversification in low and middleincome countries and increases it in high-income and all countries’ samples. Population and countries’ competitiveness are associated with increased export diversification. Additionally, innovation does not affect diversification in all the country classifications while R&D significantly promotes diversification in middle-income and reduces it in lowincome and all countries sample. The mediation effects of the variables are positive in the middle-income and highincome countries and mixed effect in low-income and all countries sample. The study, therefore recommends the need to develop human capital, increase global competitiveness, optimal use of resources for R&D to further increase export diversification in developing countries.

## 1. INTRODUCTION

Export diversification has been debated to be the key to increasing economic growth and improving the economic performance of developing countries (Agosin et al., 2012). In developing economies export diversification has been hampered by excessive trade concentration and high dependence on few primary products for export (Ibrahim and Yusuf, 2017;Newfamer et al., 2009). This dependency and the major recent events like the Covid-19 outbreak and the previous economic crisis of 2008-2009 have stirred economies around the world to diversify exports to avert external shocks necessitated by these events (Parteka, 2020). This is because any external shock may result in price shock that may adversely affect exportable commodities (Ibrahim and Yusuf, 2017). The Economic Commission of Africa (ECA) predicts the impact of the Covid-19 outbreak to be more significant in developing low-income countries, because of high dependence on primary products export (Parteka, 2020).

The degree of a country’s export diversification has been defined as the weight of different products/sectors in a country’s total export (Ibrahim and Yusuf, 2017). While in this study, export diversification is described as the ability of countries to manufacture a category of commodity compatible with the world’s export pattern. The more a country exports the higher the export diversification. Therefore, this study focuses on human capital as a knowledge-driven economy that promotes export diversification. This is important because most empirical studies have said little about the human capital effect of export diversification. It has been observed that excessive trade concentration in certain commodities is perilous for developing countries (Parteka, 2020). This has limited the gains of export diversification to developing countries.

Export diversification has many advantages such as promoting the growth of export Kehoe and Ruhl (2013), Hummels and Klenow (2005), promoting economic growth (Ibrahim and Yusuf, 2017), increasing exporter productivity Feenstra and Kee (2008), reduce the risk associated with price fluctuations (Ouedraogo et al., 2018). The impact of the diversification of export in international trade also affects the education sector because specialization in the production process will involve a country’s labour demand (Li, 2018). It is, therefore, essential to know the correct drivers of export diversification to survive external economic shocks. Melitz (2003) theoretically predicts that the reduction in variable costs can promote the export of existing products (intensive margin). However, it also provides opportunities for small companies to enter the market (extensive margin) (Töngür et al., 2020).

The index of export concentration that is known as the Herfindahl-Hirschmann Index (HHI) measures the product’s concentration level. With HHI, the more diversified a country’s export is, the smaller the concentration index. On the other hand, if the country’s concentration index approaches 1, it implies similar exportable commodities. Imbs and Wacziarg (2003) indicate that GDP has an increasing relationship with the diversification of export. Countries with higher income have a better diversification level than countries with lower income. Hummels and Klenow (2005) also added that larger economies export more product categories at specific prices with better quality. This study is based on the analysis of the determinants of export diversification in developing countries as classified according to income group.

The determinants of export diversification are very dynamic. Innovation is an essential instrument for countries to enter into new markets and increase their market share (Gunday et al., 2011). Other factors that can facilitate export diversification include decreasing uncertainty, developing human capital, and exchange rate stability (Agosin et al., 2012;Hausmann et al., 2007). The way to penetrate new export markets is through innovation and domestic market control (Cirera et al., 2015). In general, the intensity level of research and development (R&D) will impact high technology products by facilitating sectoral production capacity, improving and increasing the quality of the country’s intellectuals, increasing patent applications and innovative companies.

Most previous studies have linked export diversification to trade costs, using proxies with different logistics infrastructure (Bensassi et al., 2015;Fugazza and Hoffmann, 2017;Gani, 2017;Melitz, 2003;Töngür et al., 2020). Except for trade cost, very few studies analyzed the factors influencing the diversification of export. The novel aspect of this study is the inclusion of R&D spending and human capital as additional factors determining export diversification, the country’s competitiveness, and the country’s level of innovation. The few existing works of literature that analyzed the impact of R&D and innovation on export diversification include Zhao and Li (1997), Greenhalgh (1990). Another novel aspect of this study is that unlike previous studies Beverelli et al. (2015), Persson (2013), Dennis and Shepherd (2011) that used the count method to measure export diversification. In this study, a decomposition method is used to overcome the weakness of the count method. A study by Madzova (2018) analyzed the effect of trade competitiveness on the tendency of a country to export while the novel aspect of this study is to analyze whether competitiveness has an impact on export diversification. With this in mind, therefore, this study is important because it attempts to bridge the existing gap in the literature as stated earlier by considering the role of other factors in determining export diversification which were mostly ignored by previous studies.

The study is divided into five parts. Part 1 discusses the background of the study, part 2 discusses the theoretical basis for the study, part 3 discusses the methodology, part 4 focuses on the study findings, and part 5 summarized the study and provides suggestions for policymaking.

## 2. LITERATURE REVIEW

### 2.1 Theoretical Basis

#### 2.1.1 Export Diversification

In measuring export diversification (DE), there are many methods. The most popularly known and used by many studies is the count method. This method gives equal mass to the category of commodities and the country’s destination (Beverelli et al., 2015;Persson, 2013;Dennis and Shepherd, 2011). Another method that can be used is the decomposition method proposed by Hummels and Klenow (2005). This method can overcome the count method’s drawbacks by weighing each product with the same importance as the entire existing export category. This study used the decomposition method with export data of countries in the world taken from BACI, which were processed using the decomposition method based on Hummels and Klenow (2005). The decomposition method is expressed as follows:

$D E j t = ∑ i ε I j m t P k m i t X k m i t ∑ i ε I t P k m i t X k m i t$
(1)

where DEjt represents country j export diversification level at time t. PkmitXkmit shows the price of goods and the quantity of commodity i exports to m country at time t, the Ijmt includes categories that are observable in which country j has buoyant exports to country m at time t, k represent the country with buoyant exports to country m at time t for all I categories (in this study, I include 5,114 6-digit UN product code HS96). According to Hummels and Klenow (2005), in the calculation of the export diversification index, the whole countries in the world are used as the reference or benchmark country (k). In his case, DEjmt is the k exports to m in the category I jmt. Set in period t are related to the k exports to country m in all categories I at time t.

#### 2.1.2 Global Competitiveness Index

Krugman (1994) has defined competitiveness as the country’s ability to manufacture commodities sufficiently enough to meet the international competition standards. Meanwhile, according to the world economic forum (WEF), competitiveness is a country’s efficiency in various aspects such as institutions, policies, sustainable productive factors. The WEF uses the GCI to measure the level of competitiveness of a country. This index is very comprehensive because it has a structure that analyzes macro and micro competitiveness (Özdemİr, 2018).

#### 2.1.3 Global Innovation Index

This is an index that measures a country’s level of innovation. The index has seven indicators which are; human capital, institution, business sophistication, market sophistication, infrastructure, creative output, technology output, and knowledge. Innovation can create a differentiating effect that can be focused on capturing new market shares. Most studies at the company level show that product innovation and process innovation have a more critical role in promoting exports to companies (Becker and Egger, 2013;Cassiman et al., 2010).

#### 2.1.4 Human Capital

Martin and Orts (2001) and Ferragin and Pastore (2005) state that the production of differentiated vertically and horizontally commodities is positively determined by the level of human capital. In this study, human capital is described using the indicator of tertiary school enrolment and this is related to workers’ ability (Martin and Orts, 2001) which can increase the country’s productivity for export. Although some studies have shown a negative nexus between human capital and export diversification which mostly occurs in countries that specialized in lowskilled industries (Li, 2018;Sandu and Ciocanel, 2014).

#### 2.1.5 R&D Expenditure

R&D expenditures are expenditures on business entities, higher education, private non-profits, and government. R&D spending can be essential in the exportproductivity relationship. With reduced costs and uncertainty in entering export markets, only more productive firms can overcome entry barriers, demonstrating the importance of R&D as a stimulus for export diversification (Wakelin, 1998).

### 2.2 Empirical Review

In the case of Macedonia, Madzova (2018) shows that competitiveness can directly impact economic growth, enabling an increase in the value and volume of exports through an increased export performance in the form of a better and more diversified export structure. In another study, Bierut and Kuziemska-Pawlak (2017) explain that there is a further understanding of what is called “technological competitiveness,” which can be defined as “the capacity to innovate, as well as to increase efficiency and reduce costs.” Competitive advantage is obtained through differences in each country’s ability to innovate to provide various goods in the foreign market. Technological advancement can result in product and or process innovation. An empirical study by Osakwe et al. (2018) revealed that domestic production and export diversification plays a vital role in economic growth, even though substantial heterogeneity is seen between groups and regions of developing countries. Improved diversification is often related to the decreased volatility in output and better stability of the economy (Agosin et al., 2012). In this vein, the factors determining the diversification of export in developing countries must be looked into at the level of empirical study.

Andersson and Johansson’s (2010) study revealed that at the aggregate level, a larger city has a greater export flow. The elasticity of export flows concerning the city size variable is attributed to the extensive and intensive margins of export. The price of exported goods is negatively linked to the size of the city, while export is positively associated with the variable of city size. Furthermore, export flows’ elasticity concerning the availability of human resources was the result of export diversification adjustments. Also, the average price per unit of the different categories of the exported goods is higher in cities with higher availability of human resources. Johansson and Karlsson’s (2007) study was based on three different export diversification indicators; the exported goods, the exporting companies, and the export destinations. The empirical strategy was carried out both for total exports and for exports which were divided into ten groups of commodities. The study’s empirical findings show that high R&D intensity in manufacturing and intraregional availability of R&D companies positively influenced the three indicators of regional export diversification. Also, inter-regional availability to R&D companies has a positive and significant impact on the exported goods and the export destinations. The results of the Parteka’s (2020) study show that bilateral difference in technology as measured by the productivity of labour is the major driver of the differences in each country’s diversification of exports among the high, middle, and lowincome countries.

## 3. DATA AND METHODOLOGY

The study used panel data for 62 developing countries classified as high, middle, and low-income countries over the period 2010-2018. The data are obtained from different sources as depicted in Table 1. The most appropriate estimation technique used in this study is the Poisson Pseudo- Maximum Likelihood (PPML). The PPML technique is important in this study because in the estimate the dependent variable could be included as an index as it is in the case of our dependent variable. The PPML has the advantage of controlling for 0 observations in the data, this is important to this study as sample countries are developing with a lot of missing observations. The data in the PPML technique does not have to follow the Poisson distribution (Yadav et al., 2014). Therefore, the empirical models for the determinants of export diversification in the highincome, middle-income, and low-income countries are specified in the following nine (9) models;

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 G C I j t + μ j t$
(2)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 G I I j t + μ j t$
(3)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 R & D j t + μ j t$
(4)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 H C j t * G C I j t + μ j t$
(5)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 H C j t * G I I j t + μ j t$
(6)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 H C j t * R & D j t + μ j t$
(7)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 G C I j t * G I I j t + μ j t$
(8)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 G C I j t * R & D j t + μ j t$
(9)

$D E j t = β 0 + β 1 l n G D P P C j t + β 2 l n P O P j t + β 3 H C j t + β 4 G I I j t * R & D j t + μ j t$
(10)

Where; the subscripts j and t refer to the exporting country and time, DE is the export diversification level of the countryj, GDPPC is the per capita income in a gross term, POP is the population, HC is the human capital, GCI is the global competitiveness index, GII is the global innovation index, R & D is the R&D expenditure, HC *GCI is the interaction of human capital and country’s competitiveness, HC *GII is the interaction of human capital and country’s innovation, HC * R & D is the interaction of human capital and country’s research intensity, GCI *GII is the interactive effect of competitiveness and country‘s innovation, GCI * R & D is the interactive ef- fect of competitiveness and country’s research intensity, GII * R & D is the interactive effect of innovation and country’s research intensity and μit is the error term. The interaction terms measure the indirect effect of interacted determinants.

## 4. RESULTS

### 4.1 Descriptive Analysis

Table 2 shows the variables descriptive analysis. The Table displays the statistics based on the classification of low, middle, high, and total countries’ samples. Based on Table 2, the mean level of the diversification of export in high-income countries is the highest, which is 0.56 and exceeds all other countries’ classifications. Low-income countries have the lowest level of diversification with an index of 0.10, which is very low compared to all countries’ classifications. The average export diversification level in countries with low income is lower than the average diversification of total exports. Table 2 again display that income level possessed an increasing impact on export diversification, and all the average values of the independent variables in high-income countries are the highest in all observed country classifications.

### 4.2 PPML Estimation for Export Diversification Model

Tables 3-6 shows the empirical estimates of the export diversification model with explanatory variables in the natural logarithm which are gross domestic income per capita (lnGDPPC), natural population (lnPOP), human capital (lnHC), global competitiveness index (lnGCI), global innovation index (lnGII), spending on R&D which is in the percentage of GDP (lnR&D). The estimation results show that the critical variable i.e. human capital has varying impacts on export diversification in countries with different income levels. The estimation result of the global competitiveness index shows a significant positive impact on export diversification at a significance level of less than 1% across all models and country classifications. The global innovation index asserts no significant impact in all models and country classifications. The R&D asserts a significant impact at less than 1% level in most estimates but has a varying effect on each of the country classifications. GDP per capita and a population that were used as control variables in this study showed a significant impact at less than 1% level in most estimates.

The interaction variable between HC and the GCI shows a significant p-value in each model and classification of the observed state income towards export diversification. The interaction shows a relationship that increases the impact of the variables HC and GCI on export di-versification in developing countries. Other interaction terms show a varying impact on export diversification in the country classifications.

### 4.3 Discussion

The PPML estimation in all low, middle and the entire countries’ sample show that GDP per capita positively affects export diversification at less than 1% significance level. The positive elasticity of GDP in the low, middle, and all countries sample is shown in models 1 to 9 of Table 3, Table 4, and Table 6. The most significant positive elasticity occurs in model 3 of Table 3 i.e. 0.75% coefficient. The PPML estimation results of Table 4 also prove that the elasticity of GDP is positively associated with the diversification of export in middle-income countries. These results confirm previous studies, which also used the GDP per capita as a control variable (Beverelli et al.,2015;Dennis and Shepherd, 2011;Feenstra and Ma, 2014;Funke and Ruhwedel, 2002). The findings confirmed our a priori expectation in the low, middle, and the entire countries’ sample. The effect of GDP on diversification of export in high-income countries shows the opposite effect that is GDP reduces the diversification of export in models 1 and 4 of Table 5. The GDP's most excellent elasticity is found in model 4 of Table 5, which is 0.18. This result is supported by (Osakwe et al., 2018;Feenstra and Ma, 2014).

The PPML estimated results for all models classified according to low, middle, high, and all countries samples show a significant and positive effect of population on diversification of export. The most incredible population elasticity towards the diversification of export occurred in countries with low income (i.e. in model 2 of Table 3), which is 0.0336. Feenstra and Ma (2014) and Persson (2013) confirmed the results of this estimate. The estimate shows that the elasticity of population towards export diversification has the most significant impact in countries with low income. The elasticity of population in model 2 of Table 3 shows that an increase in the population variable by 1% increases export diversification by 1.0336%. The increase occurs because the country population average in low income is the largest of all country classifications, as shown in the descriptive statistics of Table 2.

The effect of human capital on diversification is significant in most of the estimated models, but the level of elasticity is different in the classification of the countries observed. The PPML estimated results showed a significance p-value at less than 1% level in most of the models. The effect of human capital on diversification in lowincome and middle-income countries have a negative effect. This negative effect is due to an increase in the human capital in countries with low and middle income, which focuses on low-skilled industries. This estimate supports the results of previous studies, which show that there is specialization in low-skill industries when there is an increase in human capital (Li, 2018;Sandu and Ciocanel, 2014). Human capital’s effect on export diversification in high-income countries and entire countries shows a contradicting finding. Models estimate of Table 5 show the positive influence of human capital on export diversification in high-income countries. The most outstanding elasticity is in model 1 of Table 5, which is 0.0075%. The PPML test results are also reinforced by previous studies, which observed that the effect of human capital can facilitate export diversification (Agosin et al., 2012;Andersson and Johansson, 2010;Osakwe et al., 2018;Parteka, 2020). The PPML test results are due to the accumulation of good human capital, which allows a country to change its specialization pattern concentrated on primary commodities into differentiated manufactured commodities.

The PPML estimation results in Table 3-6 show that the global competitiveness index has a significant impact at less than 1% level in all models and country classifications. The influence of the global competitiveness index on export diversification in this study shows a positive effect. In particular, the better the country’s competitiveness, the more diverse its export commodities will be. The most outstanding elasticity is found in model 1 of Table 4 in the middle-income countries. This finding confirms our theoretical expectation and is consistent with previous studies using various indicators to proxy countries’ competitiveness (Bierut and Kuziemska-Pawlak, 2017;Önsel Ekici et al., 2019;Töngür et al., 2020;Madzova, 2018).

The global innovation index does not show any significant impact across model estimates and country classifications. In the present study, the variable GII is expected to influence export diversification positively but our finding shows otherwise and thus contradicts our theoretical expectation and empirical studies by (Johansson and Karlsson, 2007;Stucki, 2016) because innovation has a different effect on export performance.

The role of the R&D expenditure variable on diversification prevailed in most of the models and country classifications. PPML estimated results of Tables 3, 4, and 6 show a significant influence of R&D spending on export diversification at less than 1% level in the low, middle, and all country classification models. The elasticity of R&D expenditures also shows a different impact on the observed country classification models. Models 3 of Tables 3 and 6 show that R&D expenditure reduces export diversification. The expenditure on R&D asserts a positive effect in model 3 of Table 4. This result indicates that the greater the expenditure on R&D in middleincome countries the higher will be export commodities diversity. The elasticity of R&D expenditure toward export diversification is 0.3617. Therefore, in the middleincome countries, the PPML estimate confirms that R&D expenditures positively affect export diversification. These results are also supported by previous studies linking R&D expenditures to export performance and export diversification (Agosin et al., 2012;Braunerhjelm and Thulin, 2008;Falk and de Lemos, 2019;Sandu and Ciocanel, 2014;Wignaraja, 2012). In high-income countries, there is no evidence of R&D impact on export diversification as shown in model 3 of Table 5. This rejects our a priori expectation that an increase in R&D expenditure will increase the level of diversification of the observed country’s exports.

The interaction between human capital and the global competitiveness index revealed a significant impact on the diversification of export in all models and country classifications at less than a 1% significance level. This variable is novel in this study because no study has ever verified the interaction of human capital with global competitiveness on the diversification of export. This finding shows that a combination of the increase in human capital and global competitiveness will strengthen the effect of the human capital and global competitiveness on the diversification of the country’s exports. Findings revealed that a 1% increase in this variable increases export diversification by 0.0018%, 0.0013%, 0.0005%, and 0.0001% in the low, middle, high, and all country income classifications. In conclusion, a country with a combination of human capital and high competitiveness will have a higher increase in export diversification.

Based on the empirical findings of Tables 3-6, human capital’s interaction with the global innovation index have only a positive and significant effect on export diversification in high-income countries. This effect is depicted in model 5 of Table 5 with an elasticity of 0.0001. These findings confirm the indirect influence of the human capital and the global innovation index that positively affects export diversification.

The empirical result shows evidence of the significant impact of human capita interaction with research and development on export diversification in low, middle, and all countries samples. Models 6 of Table 3 revealed a negative significant impact in low-income and all countries sample at less than 5% level. A unit increase in this interaction variable will reduce export diversification by 0.2089% and 0.0003% in the low-income and allcountries sample. This negative impact implies that human capital and R&D expenditure harm export diversification. That is a combination of the increase in these two variables will further reduce the level of diversification. For the middle-income, the interaction of human capital with R&D expenditure shows a positive influence on export diversification at a less than 1% significance level. The elasticity of this variable is 0.0111 which can be observed in model 6 of Table 4. This effect occurs because the variable R&D spending has a more significant impact on increasing export diversification (mode 3 of Table 4) than the negative effect on export diversification. According to model 6 of Table 5, this variable does not show any influence at all levels of significance. There is no observed evidence of the effect of the human capital with R&D expenditure interaction in countries with high income.

Model 7 of Tables 3 and 6 show no evidence of the effect of the interaction term between the GCI and the GII on export diversification. The variable shows a positive and significant effect on diversification of export in model 7 of Tables 4 and 5 at less than 10% and 5% levels. The elasticity is 0.0001, both in model 7 of Tables 4 and 5. The combination of a country’s high competitiveness, supported by a level of innovation, will increase the diversification of exports even higher.

The interaction term between the global competitiveness index and R&D expenditure on export diversification in model 8 of Tables 3 and 6 shows a negative effect at less than 5% significance level. In particular, each increase in the unit of the interaction variable of the global competitiveness index with R&D expenditures reduces the level of export diversification by 0.0193% and 0.0006% respectively. These results are contrary to the theoretical expectation. This result can occur because the negative effect of R&D expenditure (model 3 of Table 3 and model 3 of Table 6) has a more significant impact than the global competitiveness index’s effect (i.e. in model 1 of Table 3 and model 1 of Table 6). The effect of this variable is positive and consistent with our theoretical a priori in model 8 of Table 4. The effect is that a unit increase in this variable will increase export diversification of middle-income countries by 0.0059%. These results indicate that a combination of the high global competitiveness index with high R&D expenditures in middle- income countries increases export diversification even higher. There is no established evidence of a significant impact in model 8 of Table 5 in high income.

Findings revealed that the interaction between the global innovation index and the R&D expenditure has different effects in countries classification. The estimation results in model 9 of Tables 3 and 6 show a negative significant impact. This is contrary to our a priori expectation. According to the predictions, the global innovation index as interacted with R&D spending has a significant positive impact on export diversification. Specifically, a unit increase in this variable will reduce export diversification by 0.0370% and 0.0008% in the low-income and entire countries sample. The reduction occurs because R&D expenditure negatively affects low-income and entire countries (model 3 of Table and Model 3 of Table 6). The effect of this variable is positive in middle-income countries. In particular, a unit increase in this interaction term will increase the level of export diversification by 0.0073%. The estimation results of model 9 in Table 5 do not show a significant effect of this variable on export diversification in countries with high incomes.

## 5. SOME CONCLUDING REMARKS

The conclusions derived from the empirical analysis of this study are that;

• i. GDP has a positive effect on diversification of export in low, middle, and entire countries’ samples. On the other hand, it reduces diversification in countries with high income because the pattern of GDP-diversification is U-shaped. The highest elasticity is observed in countries with low income which gradually decline and turn negative in the countries with high income.

• ii. Population positively affects export diversification in low, middle, high, and all countries classification.

• iii. Human capital reduces export diversification in countries with low and middle income. The negative effect is due to the existence of export specialization in low-skilled industries. The positive influence is found in high-income and the entire countries’ sample.

• iv. The level of global competitiveness asserts a positive effect on the diversification of export in all the estimated models.

• v. The global innovation index does not affect export diversification in all the estimated models. There is no significant effect because process innovation can only increase exports in certain industries.

• vi. The R&D spending reduces export diversification in the low-income and entire countries’ sample. A positive effect is found in countries with middle income but not in countries with high income. All interaction variables have a positive impact on diversification in middle and high-income and mixed effect in low-income and all countries sample.

• vii. Human capital interacted with R&D, GCI interacted with R&D and GII interacted with R&D were found to reduce diversification in countries with low income. The interaction variable of human capital and GCI has a positive effect in countries with low income.

Consistent with the study findings, the study offered recommendations for policymakers to increase export diversification. These include;

• i. The increase in human capital is proven to increase the level of export diversification. Therefore, improving the quality of human capital can be done with equal distribution of education levels and improving the quality of education in the manufacturing sector to support export diversification.

• ii. Increasing global competitiveness by entering the market for goods that have not been penetrated before can increase the source of the country’s income. Policies that help the company export more globally can also improve the country's export patterns.

• iii. Optimal R&D spending can be of immense benefit by innovating on new products and production processes and thereby creating a comparative advantage.

• iv. A combination of improvement in human capital and competitiveness is positively associated with export diversification. The use of policies that support human capital and the country’s competitiveness will further increase export diversification.

## Table

Variables definition and sources of data

Variables descriptive analysis

Estimation results of PPML test

Estimation results of PPML test

Estimation results of PPML test

Estimation results of PPML test

## REFERENCES

1. Agosin, M. R. , Alvarez, R. , and Bravo-Ortega, C. (2012), Determinants of export diversification around the world: 1962-2000, The World Economy, 35(3), 295- 315.
2. Andersson, M. and Johansson, S. (2010), Human capital and the structure of regional export flows, Technology in Society, 32(3), 230-240.
3. Becker, S. O. and Egger, P. H. (2013), Endogenous product versus process innovation and a firm’s propensity to export, Empirical Economics, 44(1), 329-354.
4. Bensassi, S. , Márquez-Ramos, L. , Martínez-Zarzoso, I. , and Suárez-Burguet, C. (2015), Relationship between logistics infrastructure and trade: Evidence from Spanish regional exports, Transportation Research Part A: Policy and Practice, 72(2015), 47-61.
5. Beverelli, C. , Neumueller, S. , and Teh, R. (2015), Export diversification effects of the WTO trade facilitation agreement, World Development, 76, 293-310.
6. Bierut, B. K. and Kuziemska-Pawlak, K. (2017), Competitiveness and export performance of CEE countries, Eastern European Economics, 55(6), 522-542.
7. Braunerhjelm, P. and Thulin, P. (2008), Can countries create comparative advantages? R&D expenditures, high-tech exports and country size in 19 OECD countries, 1981-1999, International Economic Journal, 22(1), 95-111.
8. Cassiman, B. , Golovko, E. , and Martínez-Ros, E. (2010), Innovation, exports and productivity, International Journal of Industrial Organization, 28(4), 372-376.
9. Cirera, X. , Marin, A. , and Markwald, R. (2015), Explaining export diversification through firm innovation decisions: The case of Brazil, Research Policy, 44(10), 1962- 1973.
10. Dennis, A. and Shepherd, B. (2011), Trade facilitation and export diversification, The World Economy, 34(1), 101-122.
11. Falk, M. and de Lemos, F. F. (2019), Complementarity of R&D and productivity in SME export behavior, Journal of Business Research, 96(November 2018), 157-168.
12. Feenstra, R. and Kee, H. L. (2008), Export variety and country productivity: Estimating the monopolistic competition model with endogenous productivity, Journal of International Economics, 74(July 2008), 500-518.
13. Feenstra, R. C. and Ma, H. (2014), Trade facilitation and the extensive margin of exports, Japanese Economic Review, 65(2), 158-177.
14. Ferragina, A. M. and Pastore, F. (2005), Factor endowment and market size in EU-CEE trade: Would human capital change actual quality trade patterns? Eastern European Economics, 43(1), 5-33.
15. Fugazza, M. and Hoffmann, J. (2017), Liner shipping connectivity as determinant of trade, Journal of Shipping and Trade, 2(1), 1-18.
16. Funke, M. , and Ruhwedel, R. (2002), Export variety and export performance: Empirical evidence for the OECD countries, Weltwirtschaftliches Archiv, 138(1), 97-114.
17. Gani, A. (2017), The logistics performance effect in international trade, Asian Journal of Shipping and Logistics, 33(4), 279-288.
18. Greenhalgh, C. (1990), Innovation and trade performance in the United Kingdom, Economic Journal, 100(400), 105-118.
19. Gunday, G. , Ulusoy, G. , Kilic, K. , and Alpkan, L. (2011), Effects of innovation types on firm performance, International Journal of Production Economics, 133(2), 662-676.
20. Hausmann, R. , Hwang, J. , and Rodrik, D. (2007), What you export matters, Journal of Economic Growth, 12(1), 1-25.
21. Hummels, D. and Klenow, P. J. (2005), The variety and quality of a nation’s exports, American Economic Review, 95(3), 704-723.
22. Ibrahim, K. H. and Yuusf, M. M. (2017), Diversification of the Nigerian economy: Prospects and emerging issues in the external sector export. In A. M. Fagge, B. G. Gumel and B. A. Ahmed, Readings in Economics 2 (Diversification of the Nigerian Economy), Ahmadu Bello University Press Limited, Zaria, 284- 307.
23. Imbs, J. and Wacziarg, R. (2003), Stages of Diversification, American Economic Review, 93(1), 63-86.
24. Johansson, S. and Karlsson, C. (2007), R&D accessibility and regional export diversity, Annals of Regional Science, 41(3), 501-523.
25. Kehoe, T. J. and Ruhl, K. J. (2013), How important is the new goods margin in international trade? Journal of Political Economy, 121(2), 358-392.
26. Krugman, P. (1994), Competitiveness: A Dangerous Obsession, Foreign Affairs, 73(2), 28-44.
27. Li, B. (2018), Export expansion, skill acquisition and industry specialization: Evidence from China, Journal of International Economics, 114, 346-361.
28. Madzova, V. (2018), The impact of competitiveness on export performance of the Republic of Macedonia, Xinyang Teachers College, 10(2), 1-15.
29. Martin, J. A. and Orts, V. (2001), A two-stage analysis of monopolistic competition models of intraindustry trade, Investigaciones Economicas, XXV(2), 315- 333.
30. Melitz, M. J. (2003), The impact of trade on intra-industry reallocations and aggregate industry productivity, Econometrica, 71(6), 1695-1725.
31. Newfamer, R. , Shaw, W. , and Walkenhorst, P. (Eds.), (2009), Breaking into New Markets: Emerging Lessons for Export Diversification, World Bank Publications.
32. Önsel Ekici, Ş. , Kabak, Ö. , and Ülengin, F. (2019), Improving logistics performance by reforming the pillars of global competitiveness index, Transport Policy, 81, 197-207.
33. Osakwe, P. N. , Santos-Paulino, A. U. , and Dogan, B. (2018), Trade dependence, liberalization, and exports diversification in developing countries, Journal of African Trade, 5(1-2), 19-34.
34. Ouedraogo, R. , Sourouema, W. S. , and Zahonogo, P. (2018), Capital inflows and exports diversification in sub-saharan Africa during the MDGs Era, African Development Review, 30(1), 1-18.
35. Özdemİr, B. (2018), Effect of global competitiveness index on the export of goods and services: evidence from oecd countries, Innovation and Global Issues Congress IV, 885-891.
36. Parteka, A. (2020), What drives cross-country differences in export variety? A bilateral panel approach, Economic Modelling, 92, 48-56.
37. Persson, M. (2013), Trade facilitation and the extensive margin, Journal of International Trade and Economic Development, 22(5), 658-693.
38. Sandu, S. and Ciocanel, B. (2014), Impact of R&D and innovation on high-tech export, Procedia Economics and Finance, 15, 80-90.
39. Stucki, T. (2016), How the founders’ general and specific human capital drives export activities of start-ups, Research Policy, 45(5), 1014-1030.
40. Töngür, Ü., Türkcan, K. , and Ekmen-Özçelik, S. (2020), Logistics performance and export variety: Evidence from Turkey, Central Bank Review, 20(3), 143-154.
41. Wakelin, K. (1998), Innovation and export behaviour at the firm level, Research Policy, 26(7-8), 829-841.
42. Wignaraja, G. (2012), Innovation, learning, and exporting in China: Does R&D or a technology index matter? Journal of Asian Economics, 23(3), 224-233.
43. Yadav, M. , Jhunjhunwala, S. , Phung, Q. T. , Lupardus, P. , Tanguay, J. , Bumbaca, S. , Franci, C. , Cheung, T. K. , Fritsche, J. , Weinschenk, T. , Modrusan, Z. , Mellman, I. , Lill, J. R. , and Delamarre, L. (2014), Predicting immunogenic tumour mutations by combining mass spectrometry and exome sequencing, Nature, 515, 572- 576.
44. Zhao, H. and Li, H. (1997), R&D and export: An empirical analysis of Chinese manufacturing firms, Journal of High Technology Management Research, 8(1), 89- 105.
 Do not open for a day Close